Increasing manufacturing capabilities in a valve plant ...
Transcript of Increasing manufacturing capabilities in a valve plant ...
Faculteit Ingenieurswetenschappen
Vakgroep Technische bedrijfsvoering
Voorzitter: Prof. Dr. Ir. Hendrik Van Landeghem
Increasing manufacturing capabilities
in a valve plant
door
Nicolas Baert
Promotor: Prof. Dr. Ir. Hendrik Van Landeghem
Begeleiders: Marc Baert (Magnetrol), Veronique Limere
Masterproef ingediend tot het behalen van de academische graad van
Master in de ingenieurswetenschappen
Bedrijfskundige systeemtechnieken en operationeel onderzoek
Academiejaar 2008–2009
Faculteit Ingenieurswetenschappen
Vakgroep Technische bedrijfsvoering
Voorzitter: Prof. Dr. Ir. Hendrik Van Landeghem
Increasing manufacturing capabilities
in a valve plant
door
Nicolas Baert
Promotor: Prof. Dr. Ir. Hendrik Van Landeghem
Begeleiders: Marc Baert (Magnetrol), Veronique Limere
Masterproef ingediend tot het behalen van de academische graad van
Master in de ingenieurswetenschappen
Bedrijfskundige systeemtechnieken en operationeel onderzoek
Academiejaar 2008–2009
ACKNOWLEDGEMENTS i
Acknowledgements
Hereby I would like to show my great gratitude to everyone who contributed to the
realization of this thesis.
First of all I thank my promoter, Prof. Dr. Ir. H. Van Landeghem, who gave me the
required academic input and support for this study.
My thanks to the management of Magnetrol International NV and also its personnel
that helped me to obtain the huge amount of information and provided me with their
expertise.
Special thanks to Mr. Jeffrey Swallow, Owner of Magnetrol International Incorporated,
who gave the opportunity to work in his company.
Nicolas Baert, juni 2009
PERMISSION TO LOAN ii
Permission to Loan
“De auteur geeft de toelating deze masterproef voor consultatie beschikbaar te stellen
en delen van de masterproef te kopieren voor persoonlijk gebruik.
Elk ander gebruik valt onder de beperkingen van het auteursrecht, in het bijzonder
met betrekking tot de verplichting de bron uitdrukkelijk te vermelden bij het aanhalen
van resultaten uit deze masterproef.”
Nicolas Baert, juni 2009
Increasing manufacturingcapabilities in a valve plant
door Nicolas Baert
Masterproef ingediend tot het behalen van de academische graad van
Master in de ingenieurswetenschappen:
bedrijfskundige systeemtechnieken en operationeel onderzoek
Academiejaar 2008–2009
Promotor: Prof. Dr. Ir. Hendrik Van Landeghem
Begeleiders: Marc Baert (Magnetrol), Veronique Limere
Faculteit Ingenieurswetenschappen
Universiteit Gent
Vakgroep Technische bedrijfsvoering
Voorzitter: Prof. Dr. Ir. Hendrik Van Landeghem
Summary
This thesis includes simulations of profit and loss over a period of 10 years and isconstructed within the scope of a production overload at the Belgium facilities ofMagnetrol International. A forecast of the MINV sales is made to estimate the futureproduct mix and that product mix is projected on a 5%, 10% and 15% sales growth.Along with an analysis of production times these sales numbers where used to estimatethe future needed production capacity expressed in FTEs. Three cases were evaluated:no investment, investment and shift work. All calculations are incorporated in an Excelassessment tool. This tool is also used to calculate the influence of the US dollar/ euroexchange rate on MINV’s P&L. In order to compare the investment case with the shiftwork case a Monte Carlo simulation is performed. As a result the NPV of the cashflowsafter taxation of the investment option is higher in 60% of the cases. As a consequencethis option is preferable to the shift work option.
Key words
forecast, production capacity, profit & loss, option analysis, Monte Carlo simulation
Increasing manufacturing capabilities in a valve plant Baert Nicolas
Supervisor: Prof. dr. ir. Hendrik Van Landeghem
Abstract This article describes the methodology and the
results of the study concerning the increase of manufacturing capabilities at Magnetrol International N.V.
Keywords: forecast, production capacity, profit & loss, option analysis, Monte Carlo simulation
I. INTRODUCTION Magnetrol International N.V. is experiencing a production
overload at its plant in Zele. The overload causes the backlog and lead times to grow. This work has to convince the American owner to invest in increasing the production capabilities either by investing in new production facilities or by implementing a 2-shift system.
II. QUALITATIVE INDICATORS The increasing backlog is a direct consequence of the
overload in production, as the rate at which orders are reaching the company is larger than the rate at which they can be processed by production.
Because of the shortage of production floor space the high WIP levels cause the blocking of hallways. This makes inefficiency percentages rise and adds waste to the system.
There is no room in the current setting to separate carbon steel material production from stainless steel material production which can lead to carbon contamination of stainless steel and thus quality issues.
III. DATA ANALYSIS
A. Data collection The number of required direct production hours per product
family are acquired from the database system that keeps loggings from scanning production orders. Scanning errors were deleted by using realistic cut-off rules.
Other important data sets are the MINV accountancy system and the MINV sales reports.
.
B. Forecasting A mathematical forecast [1] is performed accompanied by a
qualitative sales survey. The survey serves to adjust those mathematical forecasts where accuracy is low because of limited historical data series. The forecasted numbers are projected on 5%, 10% and 15% growth as requested by management.
The unit sales forecast, combined with the direct production hours per product family per department and with inefficiency
N. Baert is a master of science in industrial engineering and operational
research student at the University of Ghent, [email protected].
percentages, give us the total of production hours requirement. That number is translated in the required number of full-time equivalents and in extra equipment in departments welding and assembly. The results of this section are used in the profit and loss analysis.
IV. PROFIT&LOSS ANALYSIS No investment, investment and a 2-shift system are the three
possibilities that the company wants to evaluate. To understand the P&L statement of Magnetrol International a thorough research is performed of all the statement sections and their influencing input variables. Variables like for example sales growth, inflation percentage, dollar/euro exchange rate, cost of subcontracting, cost of a new production hall and office building and so on, are considered in the analysis.
Because of the current saturation in production, making no investment does not leave any room for sales expansion. Machining and welding operations can be subcontracted but assembly forms the constraint. The cash flows in the option analysis are calculated with the no investment case as a reference.
The second possibility researched is investing in a new production and office facilities. As a third possibility a 2-shift work system has to be considered in the evaluation.
Based on the P&L analysis the investment case is preferable over the shift system because the 2-shift system is already saturated after 2014 as estimated in the 10% growth case. The P&L analysis is a very visual method but does not include the time value of money nor does it take uncertainty of input variables into account.
Figure 1. Profit after taxation
To complete the analysis an option analysis is performed.
V. OPTION ANALYSIS
A. Sensitivity analysis Multiple scenarios are imaginable with the variables set to
different values. Apart from a nominal value variables get also a high and low value to express variable variance. Through a sensitivity analysis the 4 variables with the greatest impact are selected based on the decision criterion – NPV of cash flows after taxation - and they are used to make a table with 3x3x3x3 combinations, resulting in 81 scenarios per option. The scenario probabilities are calculated from the individual probabilities of the variables’ values and a cumulative probability column is added.
B. Monte Carlo simulation By generating 10.000 random numbers - between 0 and 1 -
scenarios are selected by searching for the random numbers in the cumulative probability column. A cumulative graph is constructed with the results of this simulation.
VI. RESULTS AND CONCLUSION From Figure 2. follows that in 40% of the cases shift work
(red line on Figure 2. Monte Carlo simulation) returns a better NPV of cash flows after taxation and in 60% of the cases the investment option (blue line on Figure 2. Monte Carlo simulation). A comparison of the means also results in the investment option (€10.093.171) being preferable to the 2-shift system (€7.824.048).
In case of low sales growth it shows that the shift system is preferable. That is because there were no large investments as in the investment case that influence the NPV. Moreover the same turnover as in the investment case can be achieved in our study period of 10 years.
Figure 2. Monte Carlo simulation
ACKNOWLEDGEMENTS The author would like to acknowledge the support of Prof.
Dr. ir. H. Van Landeghem and the management of Magnetrol International N.V. to this master dissertation.
REFERENCES [1] J. Holton Wilson and Barry Keating, Business
Forecasting, McGraw-Hill Higher Education, 2002.
CONTENTS vi
Contents
Acknowledgements i
Permission to Loan ii
Overview iii
Extended abstract iv
Inhoudsopgave vi
List of Abbreviations ix
1 Magnetrol International N.V. 1
1.1 Company information . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Current situation MINV . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Need for extra capacity and floor space: indicators . . . . . . . . . . . . 4
1.3.1 Increasing backlog . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.3.2 Increase in subcontracting hours . . . . . . . . . . . . . . . . . . 5
1.3.3 Indicators on the shop floor . . . . . . . . . . . . . . . . . . . . 5
1.3.4 Safety on the shop floor . . . . . . . . . . . . . . . . . . . . . . 8
1.3.5 Quality issue . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.4 From current to future production . . . . . . . . . . . . . . . . . . . . . 10
1.5 Goal of thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2 Data Analysis 15
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.2 Direct production hours per product family . . . . . . . . . . . . . . . 15
2.3 Current production capabilities . . . . . . . . . . . . . . . . . . . . . . 19
2.4 Sales: forecasting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.4.2 Qualitative Sales Survey . . . . . . . . . . . . . . . . . . . . . . 23
2.4.3 Mathematical forecasting . . . . . . . . . . . . . . . . . . . . . . 25
2.4.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
2.5 New production capabilities . . . . . . . . . . . . . . . . . . . . . . . . 37
2.5.1 Calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
CONTENTS vii
2.5.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3 Financial Evaluation 50
3.1 Profit & Loss: introduction . . . . . . . . . . . . . . . . . . . . . . . . 50
3.2 Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
3.2.1 Economic Parameters . . . . . . . . . . . . . . . . . . . . . . . . 51
3.2.2 Profit & Loss Relations . . . . . . . . . . . . . . . . . . . . . . . 52
3.2.3 Labour Parameters . . . . . . . . . . . . . . . . . . . . . . . . . 53
3.2.4 Subcontracting Parameters . . . . . . . . . . . . . . . . . . . . . 54
3.2.5 Investment Parameters . . . . . . . . . . . . . . . . . . . . . . . 55
3.3 Assessment tool . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
3.4 Case study ’No Investment’ . . . . . . . . . . . . . . . . . . . . . . . . 56
3.4.1 Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
3.4.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
3.5 Case study ’Investment’ . . . . . . . . . . . . . . . . . . . . . . . . . . 64
3.5.1 Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
3.5.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
3.6 Case study ’Shift Work’ . . . . . . . . . . . . . . . . . . . . . . . . . . 74
3.6.1 Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
3.6.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
3.7 US Dollar analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
3.7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
3.7.2 Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
3.7.3 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
4 Option analysis 83
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
4.2 The base case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
4.3 Option analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
4.3.1 Sensitivity analysis . . . . . . . . . . . . . . . . . . . . . . . . . 85
4.3.2 Monte Carlo simulation . . . . . . . . . . . . . . . . . . . . . . 87
4.3.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
Bibliography 91
A Sales survey 92
B P&L structure 93
C Option analysis 96
D Nederlandse samenvatting 104
D.1 Inleiding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
D.2 Gegevensanalyse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
D.3 Financiele evaluatie . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
CONTENTS viii
D.3.1 Inleiding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
D.3.2 Resultaten . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
D.4 Vergelijking van de cases . . . . . . . . . . . . . . . . . . . . . . . . . . 115
D.4.1 Monte Carlo simulatie . . . . . . . . . . . . . . . . . . . . . . . 115
D.4.2 Resultaten . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
List of Figures 119
List of Tables 122
LIST OF ABBREVIATIONS ix
List of Abbreviations
BS5750 The British Standard on ’Quality Systems’.
EMEA Europe, the Middle East and Africa.
ERP Enterprise Resource Planning is a company-wide computer software
system used to manage and coordinate all the resources, information,
and functions of a business from shared data stores.
FOB Freight On Board
FTE Full Time Equivalent means a person that works full-time.
IC/PC Department for Inventory Control and Production Control
ISO International Standards Organization
MIG Metal Inert Gas welding process
MII Internal company abbreviation for Magnetrol International Inc. (USA)
MINV Internal company abbreviation for Magnetrol International N.V.
(Belgium)
MIS Management Information System
MSE Mean Squared Error
NPV Net Present Value
NS5801 Precedent of ISO 9001
OK State of Oklahoma, United States
P&L Profit and Loss statement or Income statement
PAT Profit After Taxation
PBT Profit Before Taxation
PCB Printed Circuit Board
RF Radio Frequency
RMSE Root Mean Squared Error
LIST OF ABBREVIATIONS x
ROI Return On Investment
SSE Sum of Squared Errors
TIG Tungsten Inert Gas welding process
USD/US$ The United States dollar, currency of the US. All dollars in this thesis
are US dollars.
WIP Work In Process
WTI West Texas Intermediate is a type of crude oil used as a benchmark in
oil pricing
MAGNETROL INTERNATIONAL N.V. 1
Chapter 1
Magnetrol International N.V.
1.1 Company information
The history of Magnetrol (2009) goes back to the 1930’s, when a Chicago - USA based
manufacturer of boiler systems, called Schaub Inc. was in need of level controllers for
its systems.
The level switches and controllers developed for this purpose, ultimately also became
attractive instruments for installation in applications and process installations, other
than the own Schaub systems, and gradually started to be marketed under the name
Magnetrol.
Driven by the many European projects for offshore and nuclear power plants, Magnetrol
started looking at the overseas market as a major new investment to expand business.
In 1971, Magnetrol went international, and in joint venture with the US company
Daniel Industries, founded a company called Danmag, based in Zele - Belgium. Activi-
ties of this company were sales and marketing of Magnetrol products on the European,
Middle-East and African continents, as well as manufacturing of the full range of prod-
ucts for these markets.
In 1974, Daniel Industries drew back from this joint venture, and Magnetrol became
wholly owner of Danmag. At the same time, the company was changed into Magnetrol
International N.V. Since that date, Magnetrol’s product program has been subject to
continued program and range extensions, introduction of new technologies and con-
tinued efforts for improvement of a full five year warranty on its electro-mechanical
products. Magnetrol set an innovative step forward towards supplying highly per-
forming and reliable instruments to the industry covering multiple technologies, e.g.
1.2 Current situation MINV 2
buoyancy, ultrasonic, vibrating, RF capacitance, thermal dispersion, magnetostrictive
and contact / non-contact radar.
An essential part of Magnetrol’s activities in the last decade furthermore has been the
systematic and gradual implementation of Quality Control and Quality Assurance Pro-
cedures and Programs. Such has been an essential step not only to ensure Magnetrol’s
participation in major nuclear power station projects, as well as in many offshore
projects, but also to supply quality products and services to its overall customer base.
The implemented Quality Programs resulted in Magnetrol being the first Belgian Com-
pany to have its Quality Assurance System certified in accordance with BS 5750 - part 1
and NS 5801. Certification was obtained in December, 1985 and ISO 9001 was obtained
March 1989 for the first time.
Today Magnetrol is a leading global manufacturer of industrial instrumentation, with
main markets in the oil & gas, petrochemical, power and chemical industries. The
company is headquartered in the USA and employs 600+ people worldwide.
1.2 Current situation MINV
The production environment of MINV has been developed over the years, starting at
its foundation in 1971. It is configured as a process layout where machining operations
are grouped together, as are welding and assembly operations. Today, 38 years later, it
is time to have a global look at the existing facilities and find out where improvement
in flow and layout is possible to optimize the current situation (research by Verstraete
(2009)). The reason that there has never been a radical wave of change in the pro-
duction environment of MINV, is that sales and administration are responsible for the
largest part in MINV’s cost structure and therefore they had investment priority in the
past.
It is in the scope of this thesis to consider an increase of the production capacity to
answer to the current sales boost and make a financial evaluation of it. A large part
of MINV’s sales can be assigned to customers in the energy industry, more specific the
oil companies. Due to the huge increase in oil prices over the last years as can be seen
in figure 1.1 oil companies were given the opportunity to invest more in new projects
around the world. Those investments had a direct impact on the sales of Magnetrol
International. Over the months that this thesis was written the economic situation has
changed. The crisis in the financial world has set sail to the business markets resulting
in crude oil prices below US$ 50 / barrel according to the U.S. Energy Information
1.2 Current situation MINV 3
Administration (2009).
In my research I look for a profitable way to answer to the increasing sales, and moreover
give MINV the opportunity to increase its market share in the future.
Figure 1.1: US Dollars per barrel, OK WTI spot price FOB
(Source: U.S. Energy Information Administration (2009))
In section 2.4 sales forecasts are made to harbour the capacity study of MINV’s pro-
duction facilities. A sales survey is performed in order to verify the results from the
forecast of needed production hours for the following years. An important remark is
that we don’t work with the absolute forecasted numbers, but with forecasted product
mix projections on a 5%, 10% and 15% annual growth. Information is obtained by
own research, combined with the experience and knowledge of MINV sales managers.
A proper method in this context is to make a study of the evolution at each technology
produced by MINV and make turnover forecasts.
An incoming MINV order consists of 1 or more suborders. Planning receives that order
from Factory Sales and places an order with Magnetrol International Inc. (Downers
Grove, Illinois, USA) for the needed parts. About 80% of the required parts are
procured from Magnetrol’s headquarters in Downers Grove, USA. Every Tuesday at
8 a.m. parts arrive at the production facilities of MINV in Zele, Belgium. As a
consequence of ordering parts in the USA, the reader can understand the importance
of the exchange rate study US dollar/euro. This subject attracted the attention of
Magnetrol management and they requested an analysis of the influence of the exchange
rate.
For the moment it is very favourable for MINV to order parts in the USA considering
the low exchange rate dollar/euro and the fact that special moulds require a very high
1.3 Need for extra capacity and floor space: indicators 4
investment costs that are now carried by two production facilities, in the USA and
in Belgium. Can we give a good estimate for the influence of a changing dollar/euro
exchange rate? An answer is provided in section 3.7.
Next the delivered US parts and other ordered materials/parts are kept in labelled
boxes (e.g. SM80001 5-11, Sales Magnetrol order number 80001 suborders 5 till 11)
that are stored in the stockroom.
The planning itself is done with a generic routing describing the sequential production
steps that a product must go through and the estimated production hours per process
group. Next the orders are placed in a production sequence that is discussed weekly in
a production meeting. During those meetings the production staff can decide to favour
certain orders, considering the urgency for the customer.
Thereafter a labelled box goes through the subsequent process groups: machining,
welding, assembly, quality control and shipping.
1.3 Need for extra capacity and floor space: indi-
cators
Taking a look at the company data to determine, what information there is that indi-
cates the need for production capacity expansion as well as increasing floor space.
1.3.1 Increasing backlog
Last year’s produced value was e 23,51 million. The total sales of MINV was e 26,1
million, which creates an excess of e 2,59 million. One can deduct from these numbers
that production capacity is not sufficient to fulfil the market demand. Because of
this overload the problem of long lead times arises. This is a disadvantage from a
competitive point of view. We can also see this in the increasing backlog of MINV.
If a system processes orders at a slower rate than they arrive it follows logically from
queuing theory that the queues of waiting orders will explode. Lead times will become
unacceptably high and as a consequence will negatively impact future sales as customers
request shorter lead times.
1.3 Need for extra capacity and floor space: indicators 5
1.3.2 Increase in subcontracting hours
Another indicator is the increase in subcontracting hours. Although some parts need
special expertise and are preferable for subcontracting, the main part of these subcon-
tracting hours are due to an overload in production. In other words, MINV needs to
subcontract a large amount of hours because production does not have the capacity to
produce those hours. They are on their limit.
Of course this does not necessarily mean that expensive new machines have to be
bought right away. The overload can be periodic. In this case the new machine will
have its use in some periods, but then will be idle in others.
On the other hand, in the long term this oscillation in production hours will still be
there, but the mean level of needed production hours will be higher. In that case new
machines or maybe a 2-shift system are justified.
1.3.3 Indicators on the shop floor
Have a look at figures 1.2 through 1.8 taken recently on the shop floor.
Figure 1.2: Shop floor
1.3 Need for extra capacity and floor space: indicators 6
Figure 1.3: Shop floor
Figure 1.4: Shop floor
1.3 Need for extra capacity and floor space: indicators 7
Figure 1.5: Shop floor
As can be seen on these pictures this situation brings along important safety risks. It
also slows down transport through the production hall, which increases the amount of
indirect hours and in this case can clearly be defined as waste.
Figure 1.6: Shop floor
I challenge you to manoeuvre a forklift through this hallway.
1.3 Need for extra capacity and floor space: indicators 8
Apart from the fact that there is an increasing backlog there is also a flow problem
on the shop floor. Clearly the shop cannot take extra workload as they have already
a lack of floor space at this very moment. Even if we consider extra racks, we do not
have the space to install them. Jobs that are awaiting their processing are already
blocking passage for forklifts and operators cannot move around in a safe way. Even
if an increase in production capacity was possible in the current setting, the increased
throughput would make the WIP situation only worse if cycle times of processes remain
unchanged. This follows from Little’s law:
Throughput =WIP
Cycletime(1.1)
Because of the limited production floor space there is another effect that reduces pro-
ductivity: inefficiency due to extra material handling. This effect increases the indirect
production hours and therefore the efficiency of the production force.
As you can see on figure 1.4 it is impossible to reach for the box at the back without
moving the boxes in front of it. Workers confirmed that these situations often occur.
Besides reducing work in process, racks would again be the solution. The problem is
that there is no room for more racks in the production hall.
1.3.4 Safety on the shop floor
Also note the safety advantage of extra shop floor. Now the overload of the production
facility creates possible risks for accidents as there is a lot of work in process on the
shop floor and in certain situations it blocks the way for safe passage. The handling of
longer probes in the current layout leads to inefficient material handling and could also
lead to accidents. It has not be mentioned that work accidents are a very expensive,
yet preventable, risk for the company.
1.3 Need for extra capacity and floor space: indicators 9
Figure 1.7: Manoeuvring long probe
Figure 1.8: Probe blocking safe passage
Figures 1.7 and 1.8 show workers who are trying to manoeuvre a probe of approximately
5 meter onto an assembly table and they have to stand in the painting booth to do so.
Once in place the probe doesn’t leave any room for safe passage and surely blocks any
kind of forklift or hand pallet that tries to move to the other side. In this case this
placement was only temporarily, however in many other cases with longer probes they
are blocking passage for a longer time while an assembly worker is doing a configuration
1.4 From current to future production 10
on them. For example a couple of weeks before writing this part I came across a probe
that was reaching into a welding cabin while being in configuration. Probes can go up
to 6 meter.
1.3.5 Quality issue
Last but not least: Clean/Dust areas. First of all there are some valuable technical
reasons to separate assembly from welding and machining operations. Normally it is
expected that carbon steel material production is separated from stainless steel produc-
tion, to avoid carbon contamination of stainless steel. It is also a critical requirement
to separate PCB assembly from steel production, again to avoid steel contamination
of the PCB’s. From a marketing point of view it is an advantage that a client who
visits the production plant, perceives the quality he expects from Magnetrol products
reflected in the appearance of the shop. Building a new production hall gives the
opportunity to separate assembly from welding and machining.
1.4 From current to future production
To make a profit & loss analysis of the different cases as in chapter 3 we first have to
estimate the required production capacity for the coming years. Production capacity
will be expressed in available direct hours since we deal with a job shop way of working.
This means that there are a lot of different jobs that all require a different amount of
production hours in the different departments. Most of the time the duration of a job
does not depend on the speed of a machine but on the time the worker is spending
on processing a part. In other words the processes machining, welding and assembly
within Magnetrol are all manual processes. Capacity is determined by the amount
of workers in each department. Of course a welding station cannot be occupied by
an infinite amount of people. In theory in a 1-shift system a welding station can be
used 7,5 hours per day, depending on the number of hours a worker spends on it.
At the moment all equipment (machining: machining equipment; welding: welding
station; assembly: assembly tables with configuration equipment) are continuously in
use. There is no extra capacity available.
Another remark about scalabilty of resources in the constructed model of chapters 2 and
3 is that when the amount of required direct production hours increases, the amount
of total required production hours increases with the same factor. As a consequence
the number of indirect hours also increases with the same factor. A problem with this
construction is the following: it does not necessarily imply that when we need 1000
1.4 From current to future production 11
indirect hours in the current production setting for maintenance, supervision and so on,
that we will need double of indirect hours if we want to double our production capacity.
It does not necessarily mean if we need 1 full time supervisor in the current setting
that we require 2 of them in the situation where production has doubled. These are
aspects of management control and are left out of this thesis. However, it is possible
to account for this problem in the Excel assessment tool by changing the inefficiency
reduction percentages in the parameter sheet. The same remark goes for machining and
assembly. Other supporting functions grow along with increasing sales but of course
not with the same factor (see section 3.2.3).
Now we have determined the amount of direct production hours and thus the number
of workers as a measure for the capacity, we start off with an analysis of the current
production situation. Next through sales forecasting we estimate the future require-
ment of total production hours. These numbers are converted to a required number of
workers (FTE’s) that are used in different investment options in chapter 3.
Let me first identify more specific what Magnetrol means by direct and indirect pro-
duction hours. Direct production hours are the hours that one works on a specific
production order; indirect production hours contain the period of time spent on super-
vision, quality control, maintenance of machines, material handling and safety meet-
ings. The direct production hours are measured by scanning and are logged in MINV’s
ERP system.
These data make it possible to give an estimate of future required number of FTEs
(chapter 2):
� Data analysis of database information: download in a spreadsheet and use spe-
cialized filters (pivot table as in table 2.2) to structure the information;
� Calculate average cycle times per product family per department and standard
deviation (the standard deviation can be used in case of a production simulation);
� Make a forecast of future sales numbers in units, not e values;
� Use the forecast to make a projection of the expected product mix on a 5%-10%-
15% annual sales growth;
� Use the future sales numbers in combination with the required direct hours per
product family per department to estimate the total need of direct production
hours per department over a period of 10 years.
1.4 From current to future production 12
� Convert the direct production hours to total production hours by using the in-
efficiency percentages. From the total production hours the required number of
FTEs can be calculated.
Although considering that a forecast of 10 years will generate errors of great impor-
tance, the results give us a good idea of how production will evolve in the coming years
and will be checked by performing a sales survey to increase the accuracy. The abso-
lute forecast is used to estimate the future product mix. As requested by management
this product mix is projected on the annual sales target for MINV, being a 10% sales
value growth, accompanied by a pessimistic case of 5% growth and an optimistic case
of 15%. We also assumed that a 10% sales value growth is approximately the same as
a 10% growth in unit sales.
Now we have to estimate the annually required number of full time equivalents, FTEs.
This can be done by combining the estimate of the direct production hours in combina-
tion with a most likely inefficiency percentage (share of indirect hours to the total pro-
duction hours). As mentioned above the indirect hours consist of supervision, quality
control, maintenance of machines, material handling, safety meetings and production
meetings. Part of the material handling is due to inefficiency in the current meth-
ods. In many cases this can be assigned to the lack of space in the current production
setting.
Following situations occur:
� Sometimes the material waiting for further processing is allocated in the hallway
and obstructs the passage of forklifts (space issue, inefficiency) and people (safety
issue). Indirect hours increase due to these situations.
� The racks in the production hall are not sufficient to contain all the material in
process. As a consequence of that boxes are placed on the floor, either on top of
each other or in front and next to one another. The problem here is that when
an operator has to reach for a box that is at the bottom of a pile or behind other
boxes, and then extra handling is needed to get to the box by moving others.
Again this increases inefficiency.
� Lately there is an increasing trend in probe lengths. This makes it difficult
to transport through the factory as small hallways and corners obstruct fluent
transportation of these probes. Also due to the current layout in assembly the
probes obstruct the hallways when operators are configuring them on their work
benches. Apart from increasing the inefficiency and therefore indirect hours, this
situation bring along a great safety risk.
1.4 From current to future production 13
Increasing its shop floor and changing its layout will help Magnetrol to prevent these
situations and decrease the inefficiency. In other words we cannot use the inefficiency
percentages of 2008 but we have to estimate new inefficiency percentages for the new
situation. The inefficiency percentages are used in the assessment model. The evolution
of current inefficiency percentages in the three major departments is given in the bar
chart of figure 1.9.
Figure 1.9: Inefficiency percentages in production
The combination of the future direct production hours and the inefficiency percentages
gives us the total need of production hours and therefore the number of FTEs in each
department.
Two extra production considerations:
� The maximum probe (non flexible) length is 6m. This maximum dimension is
due to transportation costs where parts longer than 6m result in extra high costs
for air transport. In the layout study we have to bring this max length into
account when considering material handling and workers’ safety.
� In the current situation electronics are assembled in the same space where steel
parts are produced. This is not an ideal situation due to contamination risks
but also customer perception. Building a new production hall would give MINV
the opportunity to make a separate ’clean room’ for the handling/assembling of
electronics.
AT MINV each year the sales targets of the previous year are multiplied by 1,1 (10%
increase) to get the targets for the next year. In the profit&loss analysis multiple values
for sales increase are considered.
1.5 Goal of thesis 14
1.5 Goal of thesis
The indicators, the large increase in sales and the overload in production show the
need for extra production floor and capacity. Increasing the capacity can be achieved
in more than one way:
� More direct production hours achieved by optimizing current facilities. By a
rearrangement of the factory layout, a better flow through the factory can be
achieved. The goal of this rearrangement is to have less transport and material
handling time so that an order will have a shorter lead time and more hours
are available to do value adding activities (e.g. more production time). For this
analysis I refer to Verstraete (2009).
There is also an opportunity to make a study of working methods and change over
times of machines, again to eliminate non-value adding activities and therefore
to create extra production capacity. This is beyond the scope of this financial
analysis.
� Shift work is another option to increase production capacity. The possibility of
2 shifts must be considered in this thesis. See also section 3.6.
� More work can be subcontracted to other companies. Since in Magnetrol’s case
subcontracting is more expensive than producing it themselves it is not a prefer-
able option.
� The production facilities can be expanded by building a new production hall
next to the existing one (room for expansion is available on MINV grounds).
This gives us also an opportunity to separate all electronic handling from steel
parts production. See also section 3.5.
In the financial P&L analysis of chapter 3 we will consider a ’no investment’, an
’investment’ and a ’shift work’ case. In chapter 4 a NPV (of cash flows after taxation)
simulation is made to compare the discounted value of the cashflows after taxation
of the two main options:implementing shift work and investing in new production
facilities.
DATA ANALYSIS 15
Chapter 2
Data Analysis
2.1 Introduction
The goal in the data analysis section is to estimate the future required production
capacity in terms of FTEs. First we determine how many direct hours are required
on the average per product family. Secondly sales numbers are forecasted to get an
estimate of the future product mix. The combination of these two results in the future
required capacity, expressed in FTEs. The outcome of this analysis is used in the
financial evaluation of chapters 3 and 4.
2.2 Direct production hours per product family
The products of Magnetrol can be divided into 10 product families:
Table 2.1: 10 Magnetrol product families
Brand name Alternative name
Mechanicals Mechanical products
Modulevel Displacer transmitters
Eclipse Guided wave radar
Pulsar Radar
Gap Sensors Ultrasonic Contact
Thermatel Thermal dispersion
Kotron RF controls
Air Sonar Ultrasonic non-contact
Jupiter Magnetostrictive
Solitel Vibrating rod
2.2 Direct production hours per product family 16
The direct hours per product family were not directly available in the MINV database
but after a little research it showed that there was a way to acquire them. MINV
stores the start and end time of a suborder in departments machining, welding and
assembly. One has to know that a Magnetrol suborder only contains products of one
type. Therefore when we know how many hours are spent on a suborder in a certain
department we can calculate how many hours are needed on the average for one product
by dividing the total time in that department(end time - start time) by the number of
products in the suborder (completed quantity).
Of course it is possible that someone works on a suborder one day and completes it
the other day. This results in two lines in the raw data list which each a start and end
time. In order to make a pivot table we have to make sure that these lines are not
summed up, as seen in following example:
First data line
Order number 79087 - 0012
Working time 2 hrs
Completed Quantity 20
Department 13 (welding)
Second data line
Order number 79087 - 0012
Working time 1,5 hrs
Completed Quantity 20
Department 13 (welding)
These two lines denote the same suborder, so the correct conclusion is that there was
3,5 hrs of work on 20 items in department 13. How can we make sure that this is also
interpreted this way when making a pivot table? There is an extra column in the raw
data list that can be used, namely: CMPLT FLAG. If this column value is 0 this order
is only partial completed, only if the column value is 1 the suborder is completed in a
department and this row only specifies the number of items completed. The working
times however must all be summed up within a suborder in a department.
Figure 2.1 shows the format of the raw data list from the MINV database.
2.2 Direct production hours per product family 17
Figure 2.1: Raw data list
Only the relevant columns are displayed. With these data we can evaluate how many
hours were spent on how many items of what type (first 3 characters of part number
denote product type) in which department.
The result of transforming the data list into a more manageable data format with the
help of pivot tables, is shown in figure 2.2.
Figure 2.2: Pivot table
As seen in figure 2.2 information is structured by: Department - Order number -
2.2 Direct production hours per product family 18
Suborder number - Product type - Sum of Production time - Maximum of Completed
Quantity.
The next step is to connect the right product codes to the correct product family. This
connection is found in the so called literature cross reference list of MINV (2009).
The numbers are transferred from the pivot table to this new list and thereafter the
total is calculated per product family in terms of total production time and completed
quantities. Subsequently we calculate the average required direct production hours per
product family and per department.
Before finalizing these numbers there was a meeting with the production manager to
correct and fine tune the data. He gave me guidelines for reasonable production times
so that I could filter the higher ones out of the analysis. The scanning system is not yet
made error proof and the situation occurs that a worker does not scan the object again
at the end of his job, which leads to high production times in the database system.
We have the sales numbers available per product family over the last 18 years and
based on these numbers we perform a forecast for the next 10 years. Those numbers
combined with the production hours per item results in the total needed production
hours in each of the departments.
Table 2.2 shows the results of the calculation described above and shows the required
direct production hours per product family item and per department.
Table 2.2: Direct production hours per product family and per department
Product family Machining Welding Assembly
Mechanical products 1,07 1,39 1,61
Displacer transmitters 1,67 2,43 2,19
Guided wave radar 1,73 1,67 1,49
Radar 0 0,31 0,96
Ultrasonic Contact 0,80 0,37 0,61
Thermal dispersion 1,42 0,33 1,19
RF controls 0 0,48 2,10
Ultrasonic non-contact 0 0,12 0,28
Magnetostrictive 0 0 0,47
Vibrating rod 0,08 0,02 0,29
Some numbers in this table may seem odd, like why does a Solitel only need 0,02 hrs
2.3 Current production capabilities 19
of welding. Normally Solitel is machined and welded in the States and the Belgium
factory assembles the products. Only in a few cases the parts coming from Downers
Grove needed some adjustment by the machining or welding departments of Zele. That
explains the low production times for this product.
Other product families like Jupiter, Air Sonar or Kotron need no machining in Belgium
at all. If Magnetrol ever decides to move the production of these items to the Belgium
factory, these numbers will have to be adjusted in the assessment tool and new time
measurements are required.
2.3 Current production capabilities
The current production capabilities are measured in production hours. The graphs
in figures 2.3 through 2.5 show us the evolution of production hours in each of the
departments from 2003 till 2008. The data to produce these graphs were deducted
from production reports of this period.
Figure 2.3: Total production hours per department
In the graph of figure 2.3 we see that assembly requires the greatest amount of total
production hours. One of the reasons assembly carries the most hours is that this
process cannot be subcontracted while machining and welding can be subcontracted.
For example in 2008 7.150 hours of welding and 2.220 hours of machining were sub-
contracted. As subcontracting is more expensive than producing the parts yourself we
try to limit this for the future, however, some parts require the expertise of extern
companies and cannot be processed at MINV.
2.3 Current production capabilities 20
Figure 2.4: Direct production hours per department
Figure 2.5: Direct production hours per department
The sudden rise of indirect production hours in 2007 is due to safety meetings and
extra material handling because of the overload in production.
Extra material handling appears in different situations, for example:
� Due to the overload in production there is a large amount of WIP. Because of
the limited floor space boxes are sometimes stacked on top of each other. When
an operator needs a box that is covered by other boxes he first has to remove
the other boxes to reach the required one. This takes lots of extra time from
operators which could be interesting to measure in the future.
� Work pieces of increasing length are also a problem on the shop floor. Recently I
took a walk through the factory and two men were busy moving long work pieces
2.4 Sales: forecasting 21
from one department to another. Due to the limited space it took quite a while
to manoeuvre them through the hallways.
2.4 Sales: forecasting
2.4.1 Introduction
A data series with a maximum of 18 historical annual figures per product is available at
MINV (2008b), which may not be sufficient to perform a forecast on with acceptable
accuracy. To prove this I perform a forecast for a given data series and then look
at the errors accompanied with this length of historical time series and the required
forecasting period. In other words, it will show that having a time series on hand of 10
years (this is the case for the Eclipse product family), the uncertainty of the forecast
for the next ten years will be quite large. This follows directly from a general wisdom
in forecasting: the further in the future you predict, the less certain your forecast will
be. A forecast is always accompanied by a forecast error, or, a forecast equals the
model value plus a certain error.
In figure 2.6 this error is graphically shown by two extra lines on a demand-time
graph. One is the upper limit with x% reliability; the other is the lower limit with x%
reliability.
Figure 2.6: Forecast illustration dr. ir. R. Van Landeghem (2008)
In the forecasts made for MINV we decided to work with the common used 90%
2.4 Sales: forecasting 22
reliability limits, being the 95% confidence level and the 5% confidence level.
These lines can be interpreted as:
� There is a 90% chance that, based on the historical time series, future demand
will have a value between the 5% confidence level and the 95% confidence level.
� There is a 5% chance that, based on the historical time series, future demand
will have a value that is lower than the 5% confidence level.
� There is a 5% chance that, based on the historical time series, future demand
will have a value that is higher than the 95% confidence level.
Let me illustrate this for the Eclipse family, number of units. Running a 10-year
forecast on the given time series from 1998 until 2008 results in the graph of figure 2.7
.
Figure 2.7: Forecast Eclipse Graph
In this picture the forecasted line, the upper and lower limit are displayed. Following
table shows the exact values.
2.4 Sales: forecasting 23
Table 2.3: Sales units forecast Eclipse
Date Annual Forecast 95% - upper 5% - lower
2009 4988 6329 3647
2010 5474 7027 3920
2011 5959 7731 4188
2012 6445 8439 4451
2013 6931 9150 4711
2014 7416 9863 4969
2015 7902 10578 5226
2016 8387 11294 5481
2017 8873 12010 5736
2018 9359 12728 5990
As you can see on figure 2.7 as well as in table 2.3, the further in the future you go, the
more uncertainty there is, or, the wider the 90% reliability lines become. In 2018 the
90% confidence interval becomes [5990,12728] produced units, which gives an interval
width of 6738 units. In 2009 the 90% confidence interval [3647, 6329] is only 2682
units.
Another measure that is used to see how well a model fits the given data is R-squared
in %. This percentage shows how well the fitted values compare to the actual values,
how well it predicts the trend in the historical time series.
For the Eclipse product family the R-squared value is 66,6%. A quick look at the
graph above shows that the model gives a very good fit for the sales numbers of 2004
and before, but cannot fit or explain the peak seen in 2007, which leads to a lower
R-squared value. For comparison, the forecast for the Thermatel Switch model has an
R-squared value of 84,3%.
2.4.2 Qualitative Sales Survey
The forecasts produced by MINV data will have a considerable error due to the limited
data. A sales survey is performed to verify the forecast results. The outcome of this
research will give us a good estimate of what the future MINV product mix will be.
The sales force can be a rich source of information about future trends and changes
in buyer behaviour. These people have daily contact with buyers and are the closest
contact the company has with its customers. It is the job of the forecaster to organize
2.4 Sales: forecasting 24
and collect this information in an objective matter to obtain a considerable insight into
future sales volumes.
This survey is sent to people of the MINV sales force to give their expectation about
the future sales of the product families in terms of five categories: highly increasing
(score=5), increase (score=4), constant (score=3), decrease (score=2), highly decreas-
ing/disappearing of the market (score=1). We will control our forecast results obtained
from a mathematical model with the product sales expectations of the sales force and
adapting the forecasted values where needed if there are large deviations between model
and sales force expectations.
Note that combining two or more forecasts is done frequently in forecasting. In our
case we are not going to add the results from one model to the results of another one,
but we are going to verify our initial forecasts with a sales survey.
The e-mail that was sent to 7 sales managers is found in Appendix A.
In reply they provided me with their sales expectations as shown in table 2.4.
Table 2.4: Sales Survey: scores as defined in text
Product family Y.D. A.R. M.B. W.H. P.S. J.V. F.A.
Mechanical products 3 2 3 4 3 3 3
Displacer transmitters 3 2 3 3 3 3 4
Guided wave radar 5 5 4 5 3 5 4
Radar 2 3 3 3 3 5 3
Ultrasonic Contact 3 3 4 3 3 2 3
Thermal dispersion 4 4 4 4 4 5 3
RF controls 1 1 2 2 3 1 1
Ultrasonic non-contact 3 2 3 3 2 2 4
Magnetostrictive 4 4 4 3 4 4 4
Vibrating rod 2 1 1 2 3 3 2
2.4 Sales: forecasting 25
Table 2.5: Summary Sales Survey
Product family Mean Standard Deviation
Mechanical products 3,00 0,58
Displacer transmitters 3,00 0,58
Guided wave radar 4,43 0,79
Radar 3,14 0,90
Ultrasonic Contact 3,00 0,58
Thermal dispersion 4,00 0,58
RF controls 1,57 0,79
Ultrasonic non-contact 2,71 0,76
Magnetostrictive 3,86 0,38
Vibrating rod 2,00 0,82
2.4.3 Mathematical forecasting
In this section the sales numbers of the 10 Magnetrol product families are forecasted
using the excel tool provided with the book Business Forecasting.
Forecasting has to be seen as a process that contains certain key components. This
process includes the selection of one or more forecasting techniques applicable to the
data that needs to be forecasted, which depends on the type of data being used.
First we have a look at the sequence one has to follow when starting a forecasting
process. As proposed in Wilson & Keating (2002):
1. Specify objectives
The objective of this forecasting is to use the future sales figures to estimate the
need of direct production hours at the production plant of MINV in Zele 10 years
from now. In this way we want to assess whether an expansion of production
capabilities in Zele is the way to go to increase profits. Apart from that it
can be interesting for Magnetrol management to see the results of an academic
forecasting based on the past sales figures.
The forecasted numbers themselves are not used to estimate the absolute quanti-
ties of products sold 10 years from now, but the relative percentages of the total.
For example it is expected that Eclipse will gain a bigger share in total sales in
the future. Because each product family hs specific required production hours
2.4 Sales: forecasting 26
this future product mix will affect the required production hours per production
department. These percentages are then projected on a 5%, 10% and 15% annual
growth.
Why do we use the forecasted numbers to estimate a product mix and then project
it? Every year, Magnetrol management sets a new target for the following year,
normally this is a sales value growth of 10%. It is this sales growth number that
management is trying to achieve every year, that will mostly determine future
sales. As worst case scenario we simulate a 5% sales growth, as best case a 15%
growth, although this second figure is quite optimistic in the current economic
setting of low economic activity. That is why the 15% growth case will get a low
probability in the option analysis of chapter 4.
2. Determine what to forecast
If one wants to know how many production hours will be needed, there has to be
a knowledge of future production demands. We have to make a forecast of future
sales and this has to be done in units, not in values, because the unit sales has a
direct influence on the required capacity. Since forecasts become more accurate
by aggregating the forecasted series, it seems a logical decision to aggregate the
different Magnetrol products to a total 10 product families (e.g. Radar, Guided
Wave Radar, Magnetostrictive, Ultrasonic Contact etc.).
3. Identify time dimensions
The length of the forecast horizon must be determined. In the case of Magnetrol,
a 10-year forecast horizon is requested to estimate the required capacity at the
Zele production plant 10 years from now. A very important note in this section
is that one has to be aware that forecasts beyond a few years are likely to be
influenced by unforeseen events that are not incorporated into the model used.
Also a forecast horizon of 10 years will generate a large error. Because of this
uncertainty concerning the forecast a qualitative forecast is made as in section
2.4.2.
We have to determine whether the forecast is required on an annual, quarterly or
monthly base. Since our goal is to estimate what production capacity is needed
over the next 10 years and expressed per year, neither quarterly nor monthly
based forecast are of any use to us. It suffices to know which capacity will
be needed from which year on. Of course this has nothing to do with the time
dimensions of the time series on hand. The more data you have, the more accurate
2.4 Sales: forecasting 27
the forecast will be, so monthly sales data would be preferable over annual data.
However we decided to work with the annual data, because retrieving monthly
data is very time consuming in the current data system.
Because we only make a one time forecast, we don’t have to pay attention to the
complexity of the forecasting methods. For example, if a forecast is meant to be
made on a very small time base (e.g. when forecasting electricity demand) then
one could prefer simple, easy to solve forecasts that not require too much time
to calculate.
4. Data considerations
In this topic one has to have a look at the quantity and type of data that are
available. The current situation is that we only have annual data available and
that we are going to try forecasting on that given time series. If it shows that
this data series gives a too large error, we might consider obtaining more detailed
information, as in monthly data. Considering the goal of forecasting in this case,
it might not be necessary to go in this much detail.
You might ask why not using the detailed information immediately? Well, this
information has to be extracted from a database in the United States and is
very time consuming, so before asking for people’s valuable time, we are going to
evaluate the time series that is directly available at MINV.
5. Model selection
After considering the objectives, what to forecast, time dimensions and data we
can go on with selecting an appropriate model. The model that will be selected
depends on several criteria:
(a) The pattern exhibited by the data
(b) The quantity of historic data available
(c) The length of the forecast horizon
We have the advantage of having a software package named ForecastX�delivered
with Wilson & Keating (2002), that has a function that through a tool called
Procast�, automatically chooses the best model to apply to the given time series.
It does this by minimizing one or more error measurements such as MSE, SSE
or RMSE.
2.4 Sales: forecasting 28
Table 2.6: Selected forecasting models
Product family Forecast model R-square
Mechanicals Holt-Winters 49,5 %
Modulevel Exponential smoothing 10,9 %
Eclipse Double exponential smoothing-Holt 66,6 %
Pulsar Exponential smoothing /
Gap Sensors Holt-Winters 83,5 %
Thermatel Switch Holt-Winters 84,3 %
Thermatel Transmitter Double exponential smoothing-Holt 84,9 %
Kotron Holt-Winters 52,5 %
Air Sonar Exponential smoothing 44,9 %
Jupiter Double exponential smoothing-Holt /
Solitel Holt-Winters 62 %
As the R-square values tell the models chosen forGap Sensors, Thermatel switch
and Thermatel transmitter sales are a good fit for the historical data of these
product families. Moduvel, Jupiter and Pulsar have a bad fit and have low or
even no R-square values were returned by the program. Having a look at figures
2.4.4 and 2.4.4 explains why a no clear trend or proper model could be found.
In the case of Pulsar there are only 6 historical sales numbers available which is
too low to make a reasonable forecast on. Also the data exhibit an unpreditable
pattern starting a low sales value, going to a very high value and then to a
medium sales value. In the case of Jupiter there is a sufficient data series (13
numbers) but with no clear pattern (long period with almost no sales).
A forecaster must be aware of the fact that using an automated forecasting
method is acceptable if you understand the selected method well enough to evalu-
ate whether it is truly a logical choice. This means that after using the automated
Procast�-function we have to check whether the chosen model seems an appro-
priate one to make a forecast on the given time series. We can check our results
with the sales expectancies of the MINV sales force, section 2.4.2.
6. Model evaluation
After selecting a workable model we have to evaluate whether it fulfils our ex-
pectancies. This can be achieved by using the sales survey and check whether
the mathematical results are in line with what people expect for the future.
2.4 Sales: forecasting 29
2.4.4 Results
Product forecasts
Air Sonar (Ultrasonic Non-Contact)
The graph in figure 2.8 below shows the forecast of the Air Sonar product family based
on historical data (blue line). The pink line shows the forecast for the next 10 years
with upper (light blue line) and lower (purple line) limits.
Figure 2.8: Forecast sales units Air Sonar
The sales survey of section 2.4.2 gave a mean score of 2,71 which indicates an expec-
tation of constant or slightly decreasing sales. This score is in line with the forecast
based on the historical data.
2.4 Sales: forecasting 30
Eclipse (Guided Wave Radar)
Figure 2.9: Forecast sales units Eclipse
The sales survey gave a mean score of 4,43 which indicates an expectation of increasing
or even highly increasing sales. This score is in line with the forecast based on the
historical data and was expected as Eclipse is Magnetrol’s best selling product at this
moment.
Gap Sensors (Ultrasonic Contact)
Figure 2.10: Forecast sales units Gap Sensors
The sales survey gave a mean score of 3 which indicates an expectation of constant
sales. This score is approximately in line with the forecast based on the historical data,
however the forecast shows a slight increase in sales, but with large uncertainty levels
as can be seen by the upper and lower limits.
2.4 Sales: forecasting 31
Jupiter (Magnetostrictive)
Figure 2.11: Forecast sales units Jupiter
The sales survey gave a mean score of 3,86 which indicates an expectation of increasing
sales. This score deviates some what from the forecast based on the historical data,
the forecast shows a highly increasing sales. Having a look at the historical data it is
clear that no clear trend can be distinguished, so a mathematical forecast may not be
that reliable here. Considering the sales survey score let us take a Jupiter sales growth
of 10% annually, parallel with the yearly sales target.
Kotron (RF controls)
Figure 2.12: Forecast sales units Kotron
The sales survey gave a mean score of 1,57 which indicates an expectation of decreasing,
maybe even highly decreasing sales. This score is not so much in line with the forecast
based on the historical data, as this forecast shows constant sales. We could make an
adaptation here. Let us take the Kotron sales evolution a 10% annual sales decrease.
2.4 Sales: forecasting 32
Mechanicals
Figure 2.13: Forecast sales units Mechanicals
The sales survey gave a mean score of 3 which indicates an expectation of constant
sales. This score is in line with the forecast based on the historical data, as this
forecast shows approximately constant sales. The forecast shows cyclic sales because
these cycles were also present in the historical data.
Modulevel (Displacer Transmitters)
Figure 2.14: Forecast sales units Modulevel
Since the best fitted forecast model only gave an R-square of 10,9 % we cannot rely
on the mathematical prediction of a constant sales. However, the sales survey gave a
mean score of 3 (indication of a constant sales expectation) and based on this finding
we assume a constant unit sales for Modulevel.
2.4 Sales: forecasting 33
Pulsar (Radar)
Figure 2.15: Forecast sales units Pulsar
The sales survey gave a mean score of 3,14 which indicates an expectation of constant
sales. As it happens, this score is in line with the forecast based on the historical data
which shows constant sales. However the forecast itself is not reliable as indicated by
the low R-square Procast�returned. The reason why Procast�could not forecast based
on the given historical data is that there was an insufficient data series of only 6 years
with no clear pattern. In this case we are relying on the sales survey that expects a
constant sales for Pulsar.
2.4 Sales: forecasting 34
Solitel (Vibrating Rod)
Figure 2.16: Forecast sales units Solitel
The sales survey gave a mean score of 2 which indicates an expectation of decreasing
sales. This score is in line with the forecast based on the historical data, as this forecast
shows a decreasing sales.
Thermatel (Thermal Dispersion switch/transmitter)
Figure 2.17: Forecast sales units Thermatel Switch
2.4 Sales: forecasting 35
Figure 2.18: Forecast sales units Thermatel Transmitter
The sales survey gave a mean score of 4 for the combination of switches and transmitters
which indicates an expectation of increasing sales. This score is in line with the forecast
based on the historical data, as this forecast shows an increasing sales trend.
Summary
The following pie charts show the estimated product mix in 2009 and 2018.
I’ve made such product mix estimation for each year from 2009 to 2018. The exact
numbers are used for a projection on a yearly 10% sales target growth, along with a
worst case of 5% and a best case of 15%.
2.4 Sales: forecasting 36
Figure 2.19: Pie chart of the forecasted product mix for 2009
Figure 2.20: Pie chart of the forecasted product mix for 2018
2.5 New production capabilities 37
2.5 New production capabilities
2.5.1 Calculations
Now that we have the required direct production hours per product family per de-
partment and an estimate of the future sales, we can proceed to calculating the new
required production capacity.
First we make a table with the actual forecasted numbers and we project these numbers
on a target growth of 10% per year. The same is done for a worst case of 5% and a
best case of 15%.
Table 2.7: Forecasted quantities 2009-2013
Forecast (qty.) 2008 2009 2010 2011 2012 2013
Mechanicals 5665 5953 6188 5398 5848 6140
Modulevel 1058 968 968 968 968 968
Eclipse 3667 4988 5474 5959 6445 6931
Pulsar 89 89 89 89 89 89
Gap sensors 2311 2407 2503 2599 2695 2791
Thermatel sw. 1107 999 1290 1293 1180 1392
Thermatel Tr. 224 249 278 307 335 364
Kotron 139 125 112 101 91 82
Air sonar 74 74 74 74 74 74
Jupiter 103 113 124 136 150 165
Solitel 79 104 139 97 109 114
Total 14515 16069 17239 17022 17985 19111
2.5 New production capabilities 38
Table 2.8: Forecasted quantities 2014-2018
Forecast (qty.) 2014 2015 2016 2017 2018
Mechanicals 6381 5566 6027 6327 6574
Modulevel 968 968 968 968 968
Eclipse 7416 7902 8387 8873 9359
Pulsar 89 89 89 89 89
Gap sensors 2887 2984 3080 3176 3272
Thermatel sw. 1497 1511 1402 1693 1696
Thermatel Tr. 393 422 451 479 508
Kotron 74 66 60 54 48
Air sonar 74 74 74 74 74
Jupiter 182 200 220 242 266
Solitel 113 60 79 104 72
Total 20075 19842 20837 22081 22927
Tables 2.7 and 2.8 show the actual forecast made by mathematical methods. To project
these numbers on the target growth (and worst and best case) we do the following.
The relationship of the actual number of a certain year to the total of that year is
calculated and multiplied by the total target sales of that year.
For example take the Mechanical sales in 2013 when a 10% sales target growth is given.
Actual Mechanicals Forecast 2013
Total Forecast 2013. T otal Target Sales 2013 =
6.140
19.111. 14.515 . 1, 12013−2008 = 7.510 (2.1)
The results are given in tables 2.9 and 2.10.
2.5 New production capabilities 39
Table 2.9: Forecast Projection on a 10% target growth 2009-2013
2009 2010 2011 2012 2013
Mechanicals 5915 6304 6127 6909 7510
Modulevel 962 987 1099 1144 1184
Eclipse 4956 5576 6763 7615 8477
Pulsar 89 91 102 106 109
Gap sensors 2391 2550 2950 3184 3414
Thermatel sw. 992 1314 1467 1394 1702
Thermatel Transm. 247 283 348 396 445
Kotron 124 114 115 107 100
Air sonar 74 76 84 88 91
Jupiter 112 126 155 177 202
Solitel 103 142 110 129 139
Total (units) 15966 17563 19319 21251 23376
Table 2.10: Forecast Projection on a 10% target growth 2014-2018
2014 2015 2016 2017 2018
Mechanicals 8173 7934 9000 9807 10795
Modulevel 1240 1380 1446 1501 1590
Eclipse 9499 11264 12524 13753 15367
Pulsar 115 128 134 139 147
Gap sensors 3698 4253 4598 4922 5373
Thermatel sw. 1918 2153 2093 2625 2785
Thermatel Transm. 503 601 673 743 834
Kotron 94 94 89 83 79
Air sonar 95 106 111 115 122
Jupiter 233 285 328 375 437
Solitel 145 86 118 161 118
Total (units) 25713 28285 31113 34225 37647
And the corresponding graph of the largest product families:
2.5 New production capabilities 40
Figure 2.21: Forecast sales of largest product families
By multiplying the projected numbers by the required direct production hours per
department the total required direct hours in each department are calculated.
With a 10% sales growth each year the needed direct hours for each department are:
Figure 2.22: Total direct hours in the machining department
2.5 New production capabilities 41
Figure 2.23: Total direct hours in the welding department
Figure 2.24: Total direct hours in the asembly department
From the bar charts in figures 2.22 through 2.24 an important conclusion to be made is
that Eclipse is the fastest grower in all three departments. This could be a justification
2.5 New production capabilities 42
to make a separate Eclipse area (machining, welding and assembly) in the new hall,
such as production manager Paul D’Hoey proposed in his latest production hall lay
out. By separating this area a lot of extra material handling can be eliminated which
increases the efficiency of the total production hall. After all the Eclipse flow is a very
important one on the shop floor and the Eclipse product family forms a great part of
the total WIP, now blocking the hallways.
The next step is to transform these direct hours back to total production hours and
full time equivalents. There are two cases when calculating the total hours from the
direct hours.
The first one is the case where there is no investment in extra shop floor and the current
inefficiencies cannot be resolved. This would be the case with the no investment and
the shift work situation. In these two situations we use the inefficiency percentages of
2008 for further calculation.
A second case is the investment in extra shop floor which makes it possible to reorganize
production to eliminate certain inefficiencies. Also the inefficiencies of extra material
handling due to a lack of space will be resolved in this way. We assume that the first
year the new hall is taken into use the inefficiency percentage will drop by 10% and
thereafter by 1% each year by performing gradual improvements. So the inefficiency
of year X is calculated as:
InefficiencyX = Inefficiency2008 . (1− 0, 1).(1− 0, 01)X−2009 (2.2)
Of course it may be possible to reduce the inefficiency even more with a very beneficial
effect, but then a detailed study of the current production system has to be made. The
reasoning is illustrated by the graphs in figures 2.25 and 2.26 for inefficiency reduction
of respectively 1%, 2% and 5% each year.
2.5 New production capabilities 43
Figure 2.25: Inefficiency reduction
Figure 2.26: Influence of inefficiency reduction on FTEs
To state this case even more the cumulative reduction in men years over 10 years
time between 1% reduction and 5% reduction (great inefficiency reduction effort) is
calculated in graph 2.27.
2.5 New production capabilities 44
Figure 2.27: Difference between 1% and 5% annual inefficiency reduction
The cumulative diffence between the 1% and 5% inefficiency reduction in the number
of FTEs over 10 years is 14 men years. The total cost of these 14 mean years, cor-
rected for the annual labour cost increase of 4%, is e 530.850 (not NPV). One can
use the assessment tool is this way to compare the benefits gained from an efficiency
improvement project against the costs of the project.
2.5.2 Results
The results of the future needed production capacity expressed in full time equivalents
is put in two graphs for each case (5%, 10%, 15% yearly sales growth). The first
graph displayed shows the future required full time equivalents in departments 12
(machining), 13 (welding) and 14 (assembly) in case Magnetrol does not invest in
a new production hall (same inefficiency situation). The second graph displays the
investment case where there is an efficiency improvement.
2.5 New production capabilities 45
Annual sales growth: 5%
Figure 2.28: Full time equivalents
Figure 2.29: Full time equivalents with better efficiency
The graphs in figures 2.28 and 2.29 show that with an investment in a new produc-
tion hall better efficiency can lead to a reduction of FTEs in machining, welding and
assembly of respectively 1, 2 and 1 in 2018. The graph below shows the cumulative
2.5 New production capabilities 46
reduction in men years in the three departments from 2009 until 2018 due to better
efficiency.
Figure 2.30: Cumulative reduction in mean years
The cumulative graph of figure 2.30 shows that over ten years time a total reduction of
8 men years in machining, 9 in welding and 8 in assembly can be achieved with better
efficiency.
Annual sales growth: 10%
Figure 2.31: Full time equivalents
2.5 New production capabilities 47
Figure 2.32: Full time equivalents with better efficiency
The graphs in figures 2.31 and 2.32 show that with an investment in a new produc-
tion hall better efficiency can lead to a reduction of FTEs in machining, welding and
assembly of respectively 3, 2 and 3 in 2018. The graph below shows the cumulative
reduction in men years in the three departments from 2009 until 2018 due to better
efficiency.
Figure 2.33: Cumulative reduction in mean years
2.5 New production capabilities 48
The cumulative graph of figure 2.33 shows that over ten years time a total reduction
of 16 men years in machining, 9 in welding and 13 in assembly can be achieved with
better efficiency.
Annual sales growth: 15%
Figure 2.34: Full time equivalents
Figure 2.35: Full time equivalents with better efficiency
2.5 New production capabilities 49
The graphs in figures 2.34 and 2.35 show that with an investment in a new produc-
tion hall better efficiency can lead to a reduction of FTEs in machining, welding and
assembly of respectively 4, 3 and 3 in 2018. The graph below shows the cumulative
reduction in men years in the three departments from 2009 until 2018 due to better
efficiency.
Figure 2.36: Cumulative reduction in mean years
The cumulative graph of figure 2.36 shows that over ten years time a total reduction
of 22 men years in machining, 15 in welding and 19 in assembly can be achieved with
better efficiency.
FINANCIAL EVALUATION 50
Chapter 3
Financial Evaluation
3.1 Profit & Loss: introduction
After the operational calculation of the required number of fulltime equivalents required
for future production, we are going to estimate the financial benefits from investing in
increased production capabilities at MINV.
Three case studies are considered:
1. no investment is made and MINV operates with its current production capacity
2. an investment is made in:
(a) doubling of the production hall
(b) extra capacity in welding and assembly, gradually
(c) extra office capacity after 5 years (beginning of 2014)
3. a 2-shift production schedule is introduced
Before calculating the future estimated P&L situation, we analyze the current structure
of MINV’s P&L. By doing this we can easily see which factors are influenced by an
increase of the manufacturing hall, the number of fulltime equivalents and production
means (extra welding units, assembly tables...). The structure of the MINV P&L is
derived from the statement Excel file available at MINV (2008a) and the result is given
in appendix B. The P&L structure is also used in chapter 4 to compare the cash flows
resulting from investing in new facilities and introducing a 2-shift system.
The next section shows the parameters that will be used in the P&L analysis and
thereafter we go to the actual P&L calculation of the ’no investment’, ’investment’ and
’shift work’ cases.
3.2 Parameters 51
The influence of the US dollar value separately on the P&L tables is shown at the end
of chapter 3 under ’US Dollar Analysis’.
3.2 Parameters
In tables 3.1 through 3.5 the parameters used for the P&L analysis are given. Not all
of the parameters are meant to be changed. Parameters like current headcount, P&L
relations, labour parameters or investment parameters are likely to remain unchanged.
Parameters like list price increase, sales growth, mean tax rate, exchange rate US
dollar/euro, increase of purchased material/parts and so on, are adjustable.
3.2.1 Economic Parameters
Table 3.1: Economic parameters
Sales growth 10 %
Increase of list prices 4 %
Estimated Future Dollar/Euro Rate 1,48
Increase of raw material prices/GOODS MII 4 %
Increase non-salary operating expenses 4 %
Mean Tax Rate 28 %
Raw material Purchase Increase 10 %
Goods purchased from MII 10 %
Increase of discount 0,50 %
Produced 2008 e 23.507.000
Exchange Rate 1e = US$ of 2008 1,48
The sales growth is the main parameter in our analysis. As displayed in table 3.1 above
is it set to 10%, but with a dropdown menu the values 5% and 15% can be chosen as
well. All P&L tables are adjusted to react in a correct way to the chosen sales growth.
For example, depending on the sales growth the limit in shift work is reached earlier
or later than in the 10% case, which had to be implemented with if-clauses.
Normally the increase of list prices depends on the Magnetrol policy. A 4% increase is
3.2 Parameters 52
now used to account for the 4% mean material price/expenses increase each year due
to price inflation.
The mean tax rate for 2008 was 28% and this value is used for later years.
The mean dollar/euro exchange rate for MINV in 2008 was 1,48 $/eFor now this value
is taken the same for the rest of the forecasted period, but can be adjusted if needed.
The influence of the US dollar/euro exchange rate is calculated in the P&L tables as
a multiplication by the 2008 exchange rate of 1,48 divided by the current rate. As
the dollar becomes more expensive this fraction becomes larger and makes MII goods
more expensive in the P&L. Each year the Magnetrol discount increases with 0,5 %
for competitive reasons explained further in this text.
3.2.2 Profit & Loss Relations
Table 3.2: Profit & Loss relations
Invoiced not shipped/List Magnetrol 0,01
Discount Magnetrol/List Magnetrol 0,27
Yearly Increase Discount 0,005
Commission/List Magnetrol 0,01
Total Other Sales/Total Invoiced 0,06
Total Other Expenses/Total Invoiced 0,05
Other Manufact Mat/Total Other Sales 0,22
The fractions or relations as calculated by dividing two 2008 P&L lines by each other.
These relations are used to calculated the lines invoiced not shipped, discount Mag-
netrol, commissions and so on, for other years.
3.2 Parameters 53
3.2.3 Labour Parameters
Table 3.3: Labour parameters
Increase services headcount 4,1 %
Increase payroll cost 4 %
Shift work salary increase 15 %
Total annual working hours in production 1706,20 hrs
Mean labour cost (no shift work) 16,50 e / h
Working days 246
Holidays 21,5
Hours per day 7,6 hrs
Salary share in total cost
Engineering 0,8565
Field Sales 0,7877
Factory Sales 0,9802
Customer Satisfaction 0,9954
Marketing 0,3813
MIS 0,7737
Administration 0,3075
Table 3.3 shows the general parameters related to labour. Most of the data were
collected from 2008 figures. To account for index adjustments of salaries there is a 4%
payroll increase each year.
The only figure here that needs some explanation is the increase in services headcount.
Management target is as follows:
� If sales increases by 5% each year or 63% over a 10 year period, then services
headcount may increase by 30% over 10 years to support the sales growth. This
equals a yearly headcount increase of 2,66%.
� If sales increases by 10% each year or 159% over a 10 year period, then services
headcount may increase by 50% over 10 years to support the sales growth. This
equals a yearly headcount increase of 4,14%.
3.2 Parameters 54
� If sales increases by 15% each year or 305% over a 10 year period, then services
headcount may increase by 70% over 10 years to support the sales growth. This
equals a yearly headcount increase of 5,45%.
3.2.4 Subcontracting Parameters
Table 3.4: Subcontracting parameters
Price Subcontract Machining 2008 40 e / h
Annual price increase subcontracting 4 %
Direct machining hours by MINV Zele Plant 11541 hrs
Subcontracting price for machining is e 40/h and is calculated from MINV accounting
numbers. The amount of direct machining hours currently available is used in further
calculations to estimate to future need for subcontracting of machining. A 4% price
inflation of subcontracting prices is incorporated.
3.2 Parameters 55
3.2.5 Investment Parameters
Table 3.5: Investment parameters
Equipment
Depreciation Period Machinery 10
Machining
Current headcount 9,5
Welding
Current numbers of welding stations 8
Relationship #welders/station 1,375
Cost of One Welding Station e 16.500
Exhausting device e 28.000
Current headcount 11
Assembly
Current headcount 12
Investment cost per head e 5.000
Maintenance cost per head per year e 800
New Buildings
Depreciation Period Building 20
Basic cost Production Hall e 1.000.000
Basic cost Administration Building e 600.000
Safety Factor 15%
Financial Cost for Loan 5%
Parameters like depreciation periods, current headcounts and investments with finan-
cial costs for building loans.
3.3 Assessment tool 56
3.3 Assessment tool
The results of section 3.4 through section 3.7 are obtained by using an Excel based
assessment tool Baert (2009) that contains all the data and calculations for the three
cases: no investment, investment and shift work. The file contains several sheets:
� a parameter sheet containing all the parameters discussed in section 3.2;
� a forecast sheet with the forecasted numbers as retrieved from ForecastX�and
the projection of these numbers on a 5%, 10% and 15% unit sales growth;
� separate sheets that calculate the required number of FTEs in case of a 5%,
10% and 15% growth. The influence of the (in)efficiency is incorporated in these
sheets;
� an investment sheet with the most important depreciations and labour costs per
year;
� separate sheets for each of the three cases:
– P&L No Investment used in section 3.4
– P&L Investment used in section 3.5
– P&L Shift Work used in section 3.6
� calculations sheets where the data of the other sheets are used and summarized to
draw conclusions concerning the dollar influence, case comparison or case analysis
(with uncertainty).
The Excel file can be found on the cd-rom accompanied with this thesis.
3.4 Case study ’No Investment’
3.4.1 Analysis
If no investment is made and we know that the current production capacity is saturated,
MINV sales can only increase if there is more work contracted out in the departments
welding and machining. This is of course without considering the assembly depart-
ment. Assembly is typically a process that cannot be subcontracted, so it remains the
bottleneck of the plant. Apart from this we can see in the present a growth of MINV
3.4 Case study ’No Investment’ 57
backlog, which will only keep on growing if no adjustments in production are made. In
other words, sales may not increase or MINV won’t be able to keep the backlog under
control.
Meanwhile costs are growing due to inflation:
� raw material gets more expensive
� labour gets more expensive (estimated 4% each year)
� subcontracting gets more expensive
MINV has to increase its sales prices or decrease its discounts in order to keep up
with these price increases and stay profitable. Of course this cannot be beneficial for
competitive reasons.
Let us begin our analysis with a definition of all the lines in the P&L of 2008 and their
relation to the input variables. For each P&L line we will estimate the future values.
In sections 3.4.1 to 3.6 all the lines are calculated using the basic values of the input
variables. In the option analysis of section 4 two or more values (low-base-high) will
be estimated for the input variables to account for uncertainty. Low-base-high values
for the input variables are agreed upon with management.
Total Net Sales
� Total Invoiced Magnetrol : sales of Magnetrol products.
– In the no-investment case List Magnetrol will be held constant, perhaps
corrected for list price changes. To be realistic we add a annual 4% price
increase which is normally taken to account for the increase in product costs,
especially payroll cost (typical increase of 4% annually). Note that this total
increase of Total Invoiced is not due to an increase in unit sales but due to
an increase of prices. Unit sales is limited by production when no investment
is made so it cannot increase.
– The production limitation has a consequence for the sales in 2009 in the
analysis. In 2008 sales was e 2,1 million too high as sales summed up to e
26,1 million and produced was only e 24 million. This became clear when
backlog rose to e 8 million where normally e 5 million is the maximum
backlog MINV can process in order to maintain a good customer service
concerning lead times. In the analysis to eliminate this excessive backlog
MINV can only sell for:
3.4 Case study ’No Investment’ 58
e 23, 507million . 1, 04− e 3million = e 21, 447million
= Total Invoiced Magnetrol allowed for 2009
Because we need to keep the partition of Total Invoiced Magnetrol into List
Magnetrol (LM, positive), Discount Magnetrol (DM, negative) and Invoiced
Not Shipped (INS, negative), we calculate them from Total Invoiced Mag-
netrol as follows:
LM2009 =21.447.000
1− DM2008LM2008
+ INS2008LM2008
.1, 04 = e 30.020.027 (3.1)
As you can see, the relationship DM/LM and INS/LM are taken from 2008.
The LM is calculated with values from 2008 and has to be corrected for a
5% price increase. Then Total Invoiced Magnetrol is calculated again with
these new figures.
Because at the end of 2009 Magnetrol will have made up arrears, 2010
values can again be calculated with the producible 2008 amount of EUR
23,507 million.
LM2010 =23.507.000
1− DM2008LM2008
+ INS2008LM2008
.1, 042 = e 34.219.611 (3.2)
The following years are calculated with the 4% price increase as before.
– Another important remark here when analyzing Total Invoiced Magnetrol
are the discounts. In general there are two kinds of discounts. One is the
discount given by direct sales people of Magnetrol to the customer. Second
is the discount that Magnetrol gives to a distributor in order to allow them
to make a profit and be competitive. When for example the distributors
ought to sell a product for 100, but cannot sell it for more then 90, they
need extra discount on the purchase price to ensure their profit margin. This
discount can go up to 32,5%.
The preceding phenomenon is due to competitors that are selling in the same
countries as the distributors, but without a distributor network and with a
direct sales force. In this way their prices can be lower than Magnetrol prices
because there isn’t a distributor in between who also needs a profit margin.
3.4 Case study ’No Investment’ 59
The distributor of Magnetrol is now forced to lower its prices if he wants to
sell his products, decreasing his margin. To help the distributors Magnetrol
grants them a standard 32,5% discount on list price. This phenomenon was
increasing in the past few years and since more than half of Magnetrol sales
is made by distributors, this effect has to be incorporated in the analysis. A
0,5% annual growth of discounts could be a realistic figure, along with 0%
and 1% as low and high value in the option analysis of chapter 4. So in the
case study we will include a constant discount and a discount increase over
10 years from 26,6% to 31,6%. This effect of uncertainty will be incorporated
in the option analysis at the end by using high - medium - low values.
Figure 3.1: The influence of discount on list prices of Magnetrol
� Total Invoiced Other : products of other companies that are distributed by Mag-
netrol. According to Magnetrol personnel, this will become zero in the future.
That is because the margins are much lower than Magnetrol products and Mag-
netrol started to produce similar products. By assumption let us decrease this
number by 1/4th of the 2008 value each year. Also correct for price increase (4%).
� Total Commissions : commissions paid to distributors or agents. Take the 2008
percentage of List Magnetrol, being 0,8%.
� Total Other Sales : this includes the sales of extra services such as packaging,
shipping, X-ray... These other sales will grow along with Total Invoiced Mag-
netrol. In 2008 Total Other Sales was 6,5% and in 2007 it was 7,44%. The mean
3.4 Case study ’No Investment’ 60
from 1993 to 2008 was 4,87%. One could conclude that there is an increasing
trend as seen in the past few years for this relationship or the mean can be used
for further calculations. In the financial analysis the mean of 4,87% is used.
Figure 3.2: Total other sales divided by total invoiced
Total Cost of Goods
� Total goods Magnetrol
– Raw Material : value of raw material purchased by the production plant in
Zele. Include 4% price increase for future years to account for inflation and
increasing prices due to world economic situation (scarcity of raw material).
– Goods MII : the dollar value of the goods bought from MII will grow along
with Total Invoiced Magnetrol. The important factor or input variable to
consider here is the exchange rate dollar/euro. The analysis in section 3.7
shows the influence of the dollar value on MINV profits. In the assessment
tool I made it is possible to change this rate in order to estimate the in-
fluence on MINV profit. In the option analysis we account for more than
3.4 Case study ’No Investment’ 61
one possible value of this rate, namely 0,9 $/e as the low value (expensive
dollar) and 1,6 $/e as the high value (cheap dollar). The value of the dollar
variable is set to 1,48 $/e in the base case.
– Inventory : inventory is counted at the end of each year. In the no-investment
case we keep this value constant, corrected for a 4% inflation. The real future
value of the inventory can differ from these numbers because of two reasons.
One is that Inventory is a snapshot of the moment and can vary depending
on the production situation at that time. Apart from this snapshot, MINV
can reduce its total inventory (WIP, stock) by introducing new techniques
into its production system such as Lean techniques.
� Total Goods Other : prices paid for products from other companies that Magnetrol
sells as a distributor (Total Invoiced Other). This value will go to 0 in a couple
of years. By assumption let us decrease this number by 1/4th of the 2008 value
each year. Also correct for inflation.
� Total Cost of Sales
– Other Manufacturing Material : this includes the costs of extra services such
as packaging, shipping, X-ray... These costs will be set equal to the 2008
percentage of Total Other Sales.
Other Manufacturing Material
Total Other Sales= 22% (3.3)
– Stock Room: depreciations of the new stock room hall, stock room forklift...
Stock room line will be held constant because the stock room hall has just
been constructed and has a depreciation period of 20 years. Apart from
that, stock room material has to be replaced and repaired from time to
time, therefore this cost too will be held constant.
– Machining : depreciations of the different machines. This value will be held
constant in the next years to account for future replacement and repair of
machines.
– Labour Machining : labour cost in Machining, increased with 4% per year.
Headcount stays the same under the no-investment decision.
– Welding : depreciations of the different machines. This value will be held
constant in the next years to account for future replacement and mainte-
nance of welding equipment.
– Labour Welding : Labour cost in Welding, increased with 4% per year. Head-
count stays the same under the no-investment decision.
3.4 Case study ’No Investment’ 62
– Assembly : depreciations of the different machines. This value will be held
constant in the next years to account for future replacement and repair of
assembly equipment.
– Labour Assembly : Labour cost in Assembly, increased with 4% per year.
Headcount stays the same under the no-investment decision.
– Quality Control : increases along with Total Invoiced Magnetrol and is in
general a payroll cost.
– Shop-General : depreciations of equipment that is not used in only one spe-
cific department, such as forklifts and production hall depreciation. The
depreciation period of the current production hall however is already com-
pleted and the only depreciations left are those of equipment that is usually
replaced after its depreciation period. All other costs in this line stay the
same under the no-investment case.
– IC/PC : this line contains only payroll costs and will be accounted for a 4%
increase in salary. Even an increase in production will not change the value
of this line. In other words, the headcount of this department will suffice
even if there is an increase in production.
– Quality Approval : for the same reason as IC/PC this line will not change,
except to account for a salary increase of 4%.
– Direct Labour Subcontracting : includes the cost of subcontracting in machin-
ing and welding. Will be held constant, except to account for an increase
in subcontracting prices of 4%. In case of no investment, will this amount
remain unchanged? If we decide not to invest, sales cannot increase because
of the limiting capacity in assembly (no subcontracting possible in assem-
bly) and with current sales subcontracting is necessary in machining and
welding to keep up with sales.
Operating Expenses - the lines written below contain not only payroll costs but also
costs like certificates (engineering), travel expenses (field sales), the mainframe de-
preciation (MIS) or the marketing budget. We separate salaries from other expenses
because salaries depend on the factor ’payroll increase’ which is 4% in the base case
and other expenses (travel, certificates, clothing, meals ...) depend on the inflation
percentage, which has also 4% as base value. In the option analysis it is possible that
both percentages differ from one another. The amount of salaries in each of the fol-
lowing service departments is calculated. These calculations are based on the numbers
found in the 2008 consolidated balance sheet and profit&loss statement of Magnetrol
Europe.
3.4 Case study ’No Investment’ 63
� Engineering
Salaries
Total Engineering=
382.742, 79
446.851, 42= 0, 8565 (3.4)
� Field Sales : from adding Sales Administration to Product Support.
Salaries
Total F ield Sales=
2.180.426, 31 + 5.987, 24
2.758.961, 16 + 16.817, 28= 0, 7877 (3.5)
� Factory Sales : from adding Export Compliance to Inside Sales.
Salaries
Total Factory Sales=
116.576, 65 + 1.226.823, 05
125.199, 56 + 1.245.338, 48= 0, 9802 (3.6)
� Customer Satisfaction
Salaries
Total Customer Satisfaction=
322.805, 47
324.282, 29= 0, 9954 (3.7)
� Marketing
Salaries
Total Marketing=
280.352, 48
735.173, 00= 0, 3813 (3.8)
� MIS
Salaries
Total MIS=
185.595, 33
239.890, 91= 0, 7737 (3.9)
� Administration: from adding Accounting to General Administration and shipping.
Salaries
Total Administration=
300.857, 24 + 215.284, 5 + 150.035, 28
367.030, 95 + 1.643.184, 96 + 156.443, 68= 0, 3075
(3.10)
Total Other Expenses - management fees for MII, lawyer expenses, financial expenses...
This value can be obtained as 5% of Total Invoiced.
3.5 Case study ’Investment’ 64
3.4.2 Results
As you are probably interested in the results of the analysis of the ’no investment’ case,
I summarized it in the graph of figure 3.3 of the estimated profit over a period of 10
years.
Figure 3.3: Financial evaluation of ’no investment’ case
3.5 Case study ’Investment’
3.5.1 Analysis
Considering the current overload in production it is clear that if management sets
increasing sales targets for the future, Magnetrol has to invest in new infrastructure to
cope with this increasing sales. We want to investigate the effect of the investments
on the P&L of MINV. In chapter 4 the effect of uncertainty of input variables is taken
into account in an option analysis.
The investments mentioned below are only a guideline and can be different from man-
agement decisions, for example the purchase of new forklifts, the expansion of certain
service departments or the installation of rolling bridges to facilitate movement of large
work pieces through the departments.
Magnetrol uses a linear depreciation system, which means the cost of a piece of equip-
ment is equally spread over the depreciation period.
3.5 Case study ’Investment’ 65
First we have to define what is understood by increasing production capabilities.
First 5 years:
� Start building a new hall in 2010 (production floor, facilities for workers) and
bring into use in 2011.
� Hiring new workers according to FTE-numbers in the assessment tool that I have
made.
� Buying new production equipment according to FTEs and reorganizing produc-
tion floor lay out.
� No expansion in machining since this is the easiest process to subcontract and
floor space is limited.
� Maybe installing rolling bridges in new production floor configuration to facilitate
movement of large and heavy work pieces through the factory.
Next 5 years:
� Take into use a new office building in 2015 along with larger demonstration room,
conference room etc. The new office building has two purposes. First of all offices
are installed in the new building when expanding manpower of the supporting
services. Secondly is the customer perception when he visits the plant.
� Hiring new workers and buying new material according to FTE-numbers in the
assessment tool Baert (2009).
� Hiring new employees to support the increase in sales.
The advantage of doing all this in phases is that management can adapt its expansion
decision whenever they believe it is appropriate. The only condition is that the lay out
is organized in such a way that it is possible to gradually install for example new welding
stations. Meanwhile the free room can be used as storage for WIP in expectation of
a production revision to shorten lead times and reduce intermediate stocks. These
two matters can evolve together: placing new equipment and finding ways to reduce
intermediate stock.
The reduction in intermediate stock will be a project for the future. Magnetrol is
considering the further elaboration of a ERP system in its production plant in Zele.
Because of the better visibility of WIP when using an ERP system compared to the
3.5 Case study ’Investment’ 66
current system, Magnetrol Zele can start working on projects to reduce WIP as well
as lead time.
Of course in the long term if their is a saturation of the new production floor shift work
can be still implemented to enable a further sales growth.
Discussion of each of the P&L lines.
Total Net Sales
� Total Invoiced Magnetrol.
In the investment case List Magnetrol will increase with a mean of 10% annually
(in the end analysis, low: 5% - high: 15%) and corrected for list price changes.
To be realistic we add a annual 4% price increase. Note that this total increase
of Total Invoiced is now due to an increase in unit sales and to an increase of
prices. Another important remark is that the expansion can be set to work until
2011, so in 2009 and 2010 production is still limited as discussed in section 3.4.1.
After 2011 sales is not limited anymore to the capacity of production as we let the
production capacity evolve in phases along with the increase in sales, as explained
before.
Thus:
– 2009 and 2010: Price adaptation, but no sales growth
– From 2011 - ...: Price adaptation and sales growth
� Total Invoiced Other : products of other companies that are distributed by Mag-
netrol. According to Magnetrol personnel, this value will evolve to zero in the
near future. That is because the margins are much lower than Magnetrol prod-
ucts and Magnetrol started to produce similar products. By assumption let’s
decrease this number by 1/4th of the 2008 value each year. Also correct for
annual price increase (4%).
� Total Commissions : commissions paid to distributors or agents. Take the 2008
percentage of List Magnetrol.
� Total Other Sales : this includes the sales of extra services such as packaging, ship-
ping, X-ray... These other sales will grow along with Total Invoiced Magnetrol.
As in the no-investment case, let us work with a mean of 4,87%.
3.5 Case study ’Investment’ 67
Total Cost of Goods
� Total Goods Magnetrol
– Raw Material : value of raw material purchased by the production plant in
Zele. Include 4% price increase every year and a 10% annual unit growth
after 2011.
– Goods MII : the dollar value of the goods bought with MII will grow along
with Total Invoiced Magnetrol. The important factor here is the exchange
rate euro-dollar that will influence MINV profits. In our analysis we will
account for several possible values of this rate.
– Inventory : increases in value with an estimated 4% per year and in amount
by 10% like the sales target growth of MINV.
� Total Goods Other : prices paid for products from other companies that
Magnetrol sells as a distributor (Total Invoiced Other). This value will evolve to
0 in a couple of years. By assumption let us decrease this number by 1/4th of
the 2008 value each year. Also correct for price increase.
� Total Cost of Sales
– Other Manufacturing Material: this includes the costs of extra services such
as packaging, shipping, X-ray... These costs will be set equal to the 2008
percentage of Total Other Sales, being 22%.
– Stock Room: depreciations of the new stock room hall, stock room forklift...
Stock room line will be held constant because the stock room hall has just
been constructed and has a depreciation period of 20 years. Apart from
that, stock room material has to be replaced and repaired from time to
time, therefore this cost too will be held constant. Increasing the MINV
production can of course lead to more material handling in the stockroom
and therefore more associated costs.
However, in line with future investments the implementation of an Inven-
tory Management system is planned. Such a system consists of processes
concerning tracking, handling and managing of goods and materials that
are held in the stockroom. After successful implementation an effective
inventory management system can not only reduce operational costs, but
reduce lead times as stock-outs are greatly reduced. The latter will increase
customer satisfaction and therefore have a beneficial effect on sales.
3.5 Case study ’Investment’ 68
– Machining : depreciations of the different machines. Since there will be no
expansion in machining this value will be held constant in the next years to
account for future replacement and repair of machines.
– Labour Machining : Labour cost in machining, increased with 4% per year.
Headcount stays the same even under the investment decision because there
will be no expansion in machining.
– Welding : depreciations of the different machines. Currently there are 11
workers (incl. foreman) in assembly for 8 welding stations. In the MINV
production presentation of 2009 we see that 6700 production hours in weld-
ing are subcontracted. This equals approximately 4 welders. This total of
15 workers very well approximates the FTEs number I calculated in the
Excel tool, which was 14. The reason these 4 extra men are contracted out
is that there is not enough room to install more welding stations.
By assumption let us use the relationship 11 workers8 welding stations
= 1, 375
We have a mean of 1,375 workers per welding station. For the future number
of welding stations we divide the forecasted number of FTEs by 1,375 and
round the result upwards. Note that we work with the 11 workers, including
foreman, because the FTEs are calculated with inefficiency percentages in-
cluding foreman. In other words, in a future FTE number there are foreman
included.
New investments include:
* New exhausting device: e 28.000 (depending on lay out)
* New welding station costs e 16.500 per station and this cost is depre-
ciated over a period of 10 years linearly.
Table 3.6: Costs welding station
1 MIG device e 6.000
1 TIG device e 6.000
1 weld manipulator e 4.500
Total cost e 16.500
Depreciation period 10 yrs
The 2008 welding cost is held constant to account for future maintenance,
repair and replacement of the current equipment.
3.5 Case study ’Investment’ 69
– Labour Welding : Labour cost in welding, increased with 4% per year. Head-
count in 2009-2010 will be held equal to the 2008 headcount because there
is no room for more workstations. From 2011-... we use the headcount cal-
culated in the assessment tool. So actually we are using the numbers of the
mentioned file with a 2-year lag.
– Assembly : depreciations of the different machines and tools. This value
will be held constant in the next years to account for future replacement
and repair of assembly equipment. From 2011 on there is room for more
equipment. Because it is not a trivial case to forecast which equipment will
be needed in the future, further research must be made. Meanwhile we will
work with estimates based on the forecasted FTEs. We calculate the total
value of the current machines and tools used by current assembly workers
and forecast these according to the number of forecasted FTEs.
Current headcount in assembly is approximately 12 according to the pro-
duction hours report of IC/PC.
Current depreciations + costs = e 20.652,64
An estimate for the investment cost per head in assembly is about e 5.000
and an extra e 800 per head per year for maintenance of equipment. The
investments can be considered per head because of the work structure within
assembly. Every blue collar has his work table with equipment, so practically
every extra worker in assembly requires a work table and specific equipment.
In other words, if there are 18 people required tomorrow in stead of 12 today
than the investment will grow by half the value it is today.
– Labour Assembly : Labour cost in the assembly department, increased with
4% per year. Headcount in 2009-2010 will be held equal to the 2008 head-
count because there is no room for more workstations. From 2011-... we
use the headcount calculated in the assessment tool. Again we are using the
numbers of the mentioned file with a 2-year lag.
3.5 Case study ’Investment’ 70
Figure 3.4: Labour costs
– Quality Control : increases along with Total Invoiced Magnetrol from 2011
on with the management target that if sales increases with 2,6 (10% annual
growth) then service personnel costs will increase with 1,5 (annual service
growth of 4,14%). This relationship is justified since further automation
will make sure that less personnel is needed for the same job. Of course the
annual payroll increase of 4% must be taken into account for all years.
– Shop-General : depreciations of material that is not used in only one specific
department, such as forklifts, production hall depreciation. In the first 2
years the same number of 2008 is used and from 2011 on following depre-
ciations are added.
* New production hall, facilities (e.g. dressing room, toilets, office):
Table 3.7: Facility costs
Price e 500/m2
Safety factor 15%
New floor surface 2.000m2
Financial loan cost 5%
Depreciation on 20 yrs e 60.375 / year
3.5 Case study ’Investment’ 71
* Expansion logistic infrastructure:
· Purchase 1 new forklift of e 11.000.
Depreciation on 10 years.
These forklifts are used to transport heavy parts.
· For lighter parts, 1 more electric pallet truck is bought for e 5.000.
* Rolling bridge (optional): 4 units across the production hall to facili-
tate the movement of big, heavy parts. This investment is left out for
now because the production lay-out are still under construction at this
moment. At the time of writing an alternative lay-out was introduced
that could eliminate the need for rolling bridges.
* Cost for rearranging shop floor.
– IC/PC : this line contains only payroll costs and will be accounted for a 4%
increase in salary. Even an increase in production will not change the value
of this line. In other words, the headcount of this department will suffice
even if there is an increase in production, according to management.
– Quality Approval : for the same reason as Quality Control this line will
increase from 2011 on with 4,14% annually. Because this is mainly a payroll
cost, we have to add 4% cost each year.
– Direct Labour Subcontracting : includes the cost of subcontracting in ma-
chining and welding, corrected for inflation. From 2011 on, since we assume
machining is not expanding, all additional hours (calculated hours - avail-
able hours) will be subcontracted. The price of a subcontracting hour is not
directly available, but is estimated to be e 40/hr.
Subsequently we multiply this cost per hour each year by the number of
hours in that year that could not be processed by the machining department
of the MINV plant. This increase will take effect from 2011 on, when the
new capacity in other departments is installed. We are going to use the
numbers forecasted in the assessment tool - 10% growth.
To calculate how many hours we have to subcontract in the future we take
the forecasted direct hours in machining and we subtract the amount of
direct hours that is currently available. These numbers are available through
production reports of 2008 that give us the total direct and indirect hours
achieved in each of the departments.
Because it was obvious from experience that in 2008 each of the departments
3.5 Case study ’Investment’ 72
worked under full capacity usage, the number obtained from these reports
is approximately the maximum amount of direct hours that can be achieved
in production at this moment, being 11.541,17 hours.
In 2009 and 2010, where the investment has not yet been completed, these
hours are held constant and the same amount as in 2008 will be subcon-
tracted because there is no growth possible due to the assembly restriction.
Because even after 2011 we assume machining department will not expand,
this department will have the same amount of direct hours as in 2008 and
every hour (as forecasted in tool) that is needed on top of this figure will be
subcontracted.
An important remark about the needed direct hours for 2008 is the following.
We see from the tool that we estimate a requirement of 17.883,8 direct
production hours in 2008 within machining. From the production reports
we can see that only 11541 hours have been achieved in Magnetrol itself plus
an extra 2.220 hours were subcontracted. The sum of these two is 13.761
hours, which does not come close to the estimated need of 17.883,8 direct
hours. Of course this phenomenon is due to the overload in production and
can be seen in the increase of the MINV backlog in 2008. In other words in
2008 Magnetrol Zele production plant could keep up with the huge increase
in sales, so, only speaking for machining, could not deliver the 17883,8 hours
needed to produce the amount that was demanded by the market.
Operating Expenses : this is a service payroll cost as well as a budget cost and a
depreciation cost and by assumption the salaries will increase with a yearly 4% and
from 2011 on an increase of service costs of 4,14%. In contrast with the no-investment
case we have to make a division of each of the lines, if relevant, because for example in
marketing the advertising budget will increase more than the payroll cost in marketing.
3.5 Case study ’Investment’ 73
The salary percentages (SP) are the same as for the no-investment case. The formulas
are:
� 2009-2010:
V alueY earX = V alueY ear2008.(SP.1, 04X−2008 + (1−SP ).1, 04X−2008) (3.11)
� 2011-...:
V alueY earX
= V alueY ear2008.(SP.1, 04X−2008 + (1− SP ).1, 04X−2008).1, 0414X−2010
(3.12)
� We apply these percentages for the lines Engineering, Field Sales, Factory Sales,
Customer Satisfaction, Marketing, MIS and Administration.
� Administration: For a good price estimate for the new administrative building
in 2015, consider the total cost of the administrative infrastructure built in 1998.
This value will be a good estimate of the cost of a new building of the same size.
The total cost is estimated to be e 600.000, plus 15% safety and 5% financial
loan cost, so e 724.500.
Total Other Expenses : management fees for MII, lawyer expenses, financial expenses...
This value can be obtained as 5% of Total Invoiced.
3.5.2 Results
The results of the analysis above are summarized in figure 3.5 of the estimated profit
over a period of 10 years. All input variables are set to their basic values. Uncertainty
of the input variables is taken into account in the option analysis of chapter 4.
3.6 Case study ’Shift Work’ 74
Figure 3.5: Financial evaluation of ’investment’ case if 10% growth occurs
3.6 Case study ’Shift Work’
3.6.1 Analysis
In case of shift work only part of the investments made in the investment case are made
such as new service personnel and a new administrative building in 2015 to handle the
increasing sales. In contrast with the no investment case shift work allows Magnetrol
to expand sales. However sales can only increase until the moment the maximum
capacity in production is reached, which will be in 2014 as mentioned below. In this
respect shift work can be considered an intermediate situation between no investment
and investment cases. If Magnetrol plans a long term growth, shift work can only be
considered as an intermediate solution, but investments have to be made to achieve
long term growth.
As a certain preparation period is necessary to set the framework for a 2-shift system,
changes will apply from 2011. Now what costs are there in a shift work environment:
Headcount
� Salary costs increase with 15% for the workers in a shift system.
3.6 Case study ’Shift Work’ 75
� Headcount will increase in each of the production departments according to the
FTEs calculated in our tool with that difference that the numbers used in the
investment case have to be corrected again. In the investment case we introduced
improvements in inefficiency because of the larger floor space. With no investment
in a larger floor space, current inefficiencies such as needless material handling,
transport and safety issues will be much harder to resolve. In other words we
keep the current inefficiency instead of correcting it for improvements such as in
the investment case. Also we could correct the FTE numbers in production for
a productivity loss due to shift work, but that is left out in this study.
Figure 3.6: Headcount in the production departments
� Shift work is also applied in machining department. Therefore less subcontracting
is necessary.
� An important remark here is that the maximum number of FTEs that can be
employed is limited. This limitation on FTEs is twice the current headcount if
there is no extra floor space created.
Current headcount is approximately (from production hours report of 2008):
– Machining : 9, 5 =⇒MAX = 19
– Welding : 11 =⇒MAX = 22
3.6 Case study ’Shift Work’ 76
– Assembly : 12 =⇒MAX = 24
A consequence is that shift work headcount can only increase until 2014 for the
10% sales growth case and therefore sales cannot increase anymore from 2015
on. If estimated headcount in machining or welding exceed the maximum we can
always subcontract this surplus at e 40/hr.
Figure 3.7: Corrected headcount in the production departments
� Also, if sales does not increase then an increase in services (engineering, field
sales, factory sales, customer satisfaction, marketing, MIS, administration) is not
necessary from 2015 on.
Up to 2014:
V alueY earX
= ServiceV alue2008.(P
T.1, 04X−2008 + (1− P
T).1, 04X−2008).1, 0414X−2010
(3.13)
Explanation: the service value is taken of the 2008 P&L and separated into a
payroll and non-payroll part, according to the P/T relationships. The payroll
3.6 Case study ’Shift Work’ 77
part is increased with 4% each year, the non-salary related costs are corrected
for inflation. Subsequently this value is increased from 2011 on, the year where
the shift work framework is completely set up, with 4,14% each year.
From 2015 on:
V alueY earX
= ServiceV alue2008.(P
T.1, 04X−2008 + (1− P
T).1, 04X−2008).1, 04142014−2010
(3.14)
Explanation: same as before, but now the service department growth is limited
to the year 2014. This is the year where the maximum capacity of assembly is
achieved and there is no production growth possible anymore.
Graph in figure 3.8 shows the P/T values of above mentioned formulas.
Figure 3.8: Operating expenses - payroll cost/total cost
The cost evolution of the different service departments is summarized in graph
3.9.
3.6 Case study ’Shift Work’ 78
Figure 3.9: Cost evolution of service departments
� To give support to each of the shifts, lead men have to be hired as well. However
we do not have to add extra costs for this as lead men are already incorporated
in the FTEs.
Subcontracting - Depending on the sales growth (5% - 10% - 15%) other situations
occur. Let’s have a look at the 10% case. Sales cannot increase from 2015 on, so the
gap in machining has to be filled up by subcontracting. To cope with the sales of 2013
and 2014 there are respectively 21 and 23 workers needed in machining and only 19
can be set to work there. So the equivalent of respectively 2 and 4 workers has to be
subcontracted, with associated costs of:
� 2 workers: 2 . 1706,2 hrs . e 40/hr . 1, 042013−2008= e 166.068
3.7 US Dollar analysis 79
� 4 workers: 4 . 1706,2 hrs . e 40/hr . 1, 042014−2008= e 345.422
Inventory
� 2009-2010: value increase of an estimated 4% per year
� 2011-2014: value increase of 4% plus amount increase of 10% per year
� 2015-...: value increase of 4% each year.
3.6.2 Results
The results of the analysis above are summarized in figure 3.6.2 of the estimated profit
over a period of 10 years. All input variables are set to their basic values. Uncertainty
of the input variables is taken into account in the option analysis of chapter 4.
Figure 3.10: Financial evaluation of the ’shift work’ case
3.7 US Dollar analysis
3.7.1 Introduction
The US dollar/euro exchange rate has known a great evolution from the beginning of
the 21st century. When the exchange rate was still under 1 before mid 2002, it is now
3.7 US Dollar analysis 80
around 1,3. In other words the dollar became cheaper over the years with regard to
the euro. Now if we zoom in on the last months there is a drop from 1,6 to 1,3 or a
change of 18,75%. One can imagine that a change of that size can have a influence
of importance on the profit of Magnetrol International N.V. which purchases a large
number of parts from its mother company in Downers Grove. Of course this influence
becomes larger as MINV purchases more parts from MII as in the ’investment’ and
’shift work’ case.
Figure 3.11: Dollar/Euro exchange rate from 2002 to 2009 BC investments (2009)
3.7.2 Analysis
Let us investigate the financial implications of the US dollar/euro exchange rate for
MINV in the three cases we’ve been studying in the beginning of this chapter. As
studied rate values let us consider a rate of 1,28 around the time this analysis was
written, 1,48 as the mean rate of 2008 that was established in the MINV accounts and
a rate of 1,6 as was the peak in the summer of 2008. To be complete a dollar/euro rate
of 0,9 is added to the analysis. 0,9 $/e was the exchange rate only seven years ago.
The impact of these changes are easily calculated with the Excel assessment tool. The
analysis below goes out from the assumption that the chosen rate is the mean rate
throughout the total period studied, which is 10 years in this thesis.
For the case where no investment is made the influence on the profit (after tax) of the
US dollar/euro exchange rates is shown graphically in figure 3.12.
3.7 US Dollar analysis 81
Figure 3.12: Influence of dollar/euro exchange rate on the profit in the ’no investment’ case
The difference in profit between the 0,9 rate and the 1,6 rate in 2009 is e 2.573.489
and in 2018 e 3.662.878.
For the case where the investment is made the influence of the US dollar/euro exchange
rates is shown graphically below.
Figure 3.13: Influence of dollar/euro exchange rate on the profit in the ’investment’ case
The difference in profit between the 0,9 rate and the 1,6 rate in 2009 is e 2.573.489
and in 2018 e 7.851.704.
For the case where a shift work system is being introduced the influence of the US
dollar/euro exchange rates is shown graphically in figure 3.14.
3.7 US Dollar analysis 82
Figure 3.14: Influence of dollar/euro exchange rate on the profit in the ’shift work’ case
The difference in profit between the 1,28 rate and the 1,6 rate in 2009 is e 2.573.489
and in 2018 e 5.362.820.
3.7.3 Conclusion
Apparently the US dollar/euro rate is a factor that cannot be neglected. In all three
cases a shift in long term US dollar/euro exchange rate can have a large influence on
the profit of MINV since a large amount of parts is purchased in the USA. Having a
look at the P&L tables in the assessment tool shows clearly the relative importance of
the GOODS MII line.
OPTION ANALYSIS 83
Chapter 4
Option analysis
4.1 Introduction
A rational way to compare 2-shift system with the investment case is to make an
option analysis of the situation. In an option analysis different options or cases are
analyzed using simulation techniques in order to make allowance for uncertainty. A
decision has to be made which option to implement based on a decision criterion. Each
option can consist of multiple decisions, but these underlying decisions are bundled in
two options: investment (investing in larger production hall, new office building, new
production equipment) and shift work (introducing a 2-shift system, investing in new
office building).
In this thesis there are in fact three cases, options: no investment, investment and shift
work. The cashflows of the investment case and shift work case are calculated relative
to the no investment option. In this way the direct gain from the option can be seen.
Next determine which categories are deterministic and are calculated from other cate-
gories (appendix B). These categories form the deterministic model and will be calcu-
lated from the ones that are not deterministic. Examples of these categories are Total
Invoiced, Total Net Sales, Commissions...
All other fields containing for example labour costs, material costs, subcontracting
costs and cost of goods from the USA (due to uncertain dollar value), are probabilistic.
There is uncertainty about their values because they depend on input variables that
are hard to predict. In this case we will make assumptions and give a range of 2 or 3
values for the input variables. For example, for the dollar/euro exchange rate 3 possible
values are included: 0,9 $/e - 1,48 $/e - 1,6 $/eThe exchange rate affects the Euro
price of the goods bought from the production unit in Downers Grove. Apart from a
4.2 The base case 84
nominal value, we have a high and a low value to represent the variation of the variable
due to uncertainty.
The uncertainties as mentioned above are used in the Monte Carlo simulation of section
4.3.2 in which only the input variables with the most impact on the NPV of the cash
flows, can vary. The impact of the input variables can be calculated by performing
a sensitivity analysis where for each variable the decision criterion is calculated for
the different variable values and the other variables are kept at their base value. The
variables that cause the greatest change in the NPV are used in the Monte Carlo
simulation.
4.2 The base case
Graph 4.2 shows a comparison of the P&L profits after taxation for the three cases
where the input variables are set to their basic values as in section 3.2.
Figure 4.1: PAT over a period of 10 years for the 3 cases
There are two reasons why this analysis is too superficial. One is that all the input
variables’ values are set to their basic values which excludes the effect of uncertainty.
For example, the investment cost of a new production facility is set to e 1.000.000,
but depending on the degree of finishing it could easily go up to e 1.500.000. Another
example is that the dollar/euro exchange rate is set to 1,48 $/e but the past showed
rates below 1 $/e or even up to 1,6 $/e .
4.3 Option analysis 85
The second reason why this analysis does not suffice is that the true impact of the
investments is not correctly implemented in the P&L calculations. First of all the time
value of money is neglected. There is a difference in spending e 1.000.000 in 2010 or
spending it in 2014. Therefore we should work with discounted values as calculated
later on. Also the P&L analysis shows no direct influence from the money invested,
only from the depreciations resulting from the investments.
For these reasons we proceed to a so called option analysis where we use the net present
value of the cash flows as a reliable measure for the profitability of investment and shift
work case.
4.3 Option analysis
Uncertainty makes it unreliable to compare investment and shift work case with input
variables set to their mean values. A more profound study can be found in the option
analysis. The goal of such an analysis is to compare to possible options in case of
uncertainty.
First an influence diagram is made such as in appendix B to understand the structure
and be able to perform the calculations. All of these relations are incorporated in the
Excel assessment tool resulting from the detailed study in the beginning of this chapter.
4.3.1 Sensitivity analysis
The variation of some input variables in the assessment tool are subjected to uncer-
tainty and those variables are tested in a sensitivity analysis where a variable is given 2
or 3 possible values: a nominal or basic value, a low and high value. The variables that
give the largest deviation in the decision criterion, NPV of cash flows after taxation,
are held for further analysis. The following variables are considered as being uncertain
and are used for a sensitivity analysis.
4.3 Option analysis 86
Table 4.1: Chosen variables for a sensitivity analysis
Investment Shift work Low - Base - High
Sales growth x x 5% - 10% - 15%
Dollar value x x 0,9 - 1,48 - 1,6 $/e
Inflation percentage x x 3% - 4% - 5%
Cost production hall x e 1.000.000 - e 1.500.000
Cost administrative building x x e 600.000 - e 900.000
Payroll increase x x 2% - 4%
Discount increase x x 0% - 0,5% - 1%
Initial ineff. decrease x 5% - 10% - 15%
Yearly ineff. decrease x 0% - 1% - 2%
Subcontracting hour x x e 35 - e 40 - e 50
The timing of the building investments is fixed (2010 and 2014). The timing of the
investments in welding and assembly depend on the sales growth that is experienced
at that time and is incorporated in the calculations of the assessment tool.
Net present value of cash flows after taxation is chosen as the decision criterion and is
calculated as
9∑i=1
(1− Tax rate).savingsi + Tax rate.depreciationi − capital expenditurei
(1 + discount rate)i(4.1)
Tables C.1 and C.2 of appendix C show the cash flows resulting from each option
(investment or shift work) for every year, starting with 2009 (year 0) till 2018 (year
9). This can be achieved by subtracting the cash flows of the no investment case from
the option under analysis resulting in the cash flows only depending on the either the
investment decision or the shift work decision. The cash flows are deducted from the
P&L tables. The tax savings from the new depreciations are calculated separately and
are also displayed in the tables. The value of the capital investments are displayed in
the upper line. At the bottom of each table the NPV of the cash flows after tax are
displayed. Note that in the cash flow table for the investment option 2009 is left out
and in the shift work case 2009 as well as 2010 because all values are 0.
Next we determine which input variables are to be used in the Monte Carlo simulation.
To determine this a sensitivity analysis is performed for the variables of table 4.3.1.
4.3 Option analysis 87
The results of this analysis can be found in tables C.3 and C.4 of appendix C. From
each table we select the 4 input variables that give the largest deviation on the NPV,
being: sales growth, dollar value, inflation percentage and yearly discount increase.
These variables are selected for the Monte Carlo simulation.
The results of the sensitivity analysis are as follows. The base value for option 1 is e
13,855 million and e 10,155 million for option 2. The value ranges are e 4,522 million
- e 25,831 million for option 1 and e 5,575 million - e 14,161 million for option 2. As
the reader can see both intervals overlap and no clear conclusion can be drawn.
4.3.2 Monte Carlo simulation
The next step is performing a Monte Carlo simulation. First a table is constructed with
all possible combinations of the 4 selected input variables with 3 values each, resulting
in 81 possible scenarios per option. The combinations are displayed in tables on pages
96 and 98 of appendix C. The probabilities of each of the variables’ values are given in
table 4.3.2. By multiplying the correct probabilities (see table 4.3.2) for each scenario
the probability is calculated that a certain scenario occurs. As a preparation for the
Monte Carlo simulation an extra column is added with the cumulative probabilities.
Table 4.2: Probabilities of variables’ values
Variable Low Base High
Sales growth 0,45 0,45 0,1
Dollar value 0,33.. 0,33.. 0,33..
Inflation percentage 0,4 0,4 0,2
Yearly discount increase 0,25 0,5 0,25
Before proceding take a quick look at the 3x3x3x3 combination tables on pages 96 and
98. Clearly the shift work option is better than the investment option if there is a
low sales growth (5%). In case of a large sales growth (10% and 15%) the investment
option is preferable.
After completing the scenario tables a new Excel table is constructed with 10.000
random numbers generated for each option. The random numbers are generated by the
Excel function Rand() between 0 and 1, and are meant to make a random selection from
the scenario/combinations list. For each random number Excel searches its position in
the cumulative probability column of the scenario list. If the exact number is not in the
column then the greatest number smaller than our number is selected. As output the
4.3 Option analysis 88
search function gives the NPV value in the row below the number that was found. That
is because Excel finds the lower limit of the interval that contains the random number
and we need the upper limit of the interval. An easy way to understand this way of
simulating is to see it as a darts game. Each scenario gets a part of the disc according
to its probability where the sum of all the slices adds up to the whole disc (100 %). A
blindfolded person throws a dart at the disc of figure 4.3.2 and hits randomly one of
the slices. This will be the selected scenario. The larger a slice is (or the larger the
probability of that scenario is), the more chance it has to get hit by the dart.
Figure 4.2: Darts game illustration of Monte Carlo simulation
This is done for all 10.000 random numbers and for each option, and a mean is calcu-
lated per option. In other words Excel simulates 10.000 times a possible scenario where
scenarios with a higher probability are more likely to occur. The mean of all 10.000
NPV values is calculated for each option. Each time Excel simulates 10.000 random
numbers this mean can differ a bit. To conclude the analysis an extra table is made
with 200 of such means for both the investment as the shift work option. Finally the
mean of the means is calculated. As can be easily understood these two numbers are
more reliable to compare with each other than taking the NPV values for the base case
and comparing those values. By performing a Monte Carlo simulation uncertainty is
taken into account and more reliable conclusions can be drawn.
4.3 Option analysis 89
4.3.3 Results
The Monte Carlo simulations themselves are too long to include on paper but are
supplied on the cd-rom delivered with this book. The results of the simulation are
displayed in the cumulative graph of figure 4.3.3.
Figure 4.3: Cumulative probabilities as perceived in Monte Carlo simulation
Based on graph 4.3.3 we can draw the conclusion that in 40% of the cases the shift
work option is preferable to the investment option. In 60% of the cases the investment
option is better. The horizontal distance between the graphs denote the difference in
NPV. Of course this analysis cannot predict what will really happen, but it gives a
good idea of the difference between the shift work and investment option.
In 30% of the cases the NPV of the cash flows resulting from investing is below e
5.000.000 and in less than 20% of the cases the NPV resulting from introducing a
2-shift system is less than e 5.000.000. The scenarios where a shift system performs
better are in the lower end tail of the distribution and have typically a low sales growth
of 5 %. If Magnetrol is expecting a low sales growth then implementing a 2-shift system
might be the better solution (when not considering the limited floor space issues).
In 10% of the cases the NPV of the shift work option is above e 11.000.000 and in
45% the NPV of the investment option is above this value. In the higher end tail the
investment option shows to be the better solution. That is because the 2-shift option
is quickly saturated which puts a stop to further sales growth.
4.3 Option analysis 90
Taking the mean of the simulation of 10.000 numbers and simulating this mean 200
times gives a mean of means of e 10.093.171 for the investment option and e 7.824.048
for the shift work option, in favour of the investment option.
Based on the cumulative graph and the calculated means, we conclude that invest-
ing in new production facilities is the better option of the two. Besides the financial
conclusion, one must also consider the social impact of the 2-shift option. First man-
agement has to make the current workers agree to work in shifts, and secondly the
immense difficulties have to be considered to find additional qualified workers that are
willing to work in shift. One has to consider that in the industrial park of Zele only
few companies work in shift, so that finding workers in the area to do so may be very
difficult if not impossible.
BIBLIOGRAPHY 91
Bibliography
N. Baert (2009). Assessment tool. Assessment tool.xls.
BC investments (2009). Dollar/Euro exchange rate. http://www.bcinvestments.
net/chart.php.
P. dr. ir. R. Van Landeghem (2008). Advprolog forecasting 2008 x. ADVPRO-
LOG forecasting 2008 X.pdf.
Magnetrol (2009). Magnetrol International - About Us. http://www.magnetrol.com/
uk/html/about_us.asp.
MINV (2008a). Balance 12-07 auditor. BALANCE 12-07 AUDITOR.xls.
MINV (2008b). Statistics MINV November 2008 YTD. Statistics MINV2008NOV.xls.
MINV (2009). Literature cross reference list. Literature cross reference list.xls.
U.S. Energy Information Administration (2009). Cushing, OK WTI Spot Price FOB
(Dollars per Barrel). http://tonto.eia.doe.gov/dnav/pet/hist/rwtcd.htm.
C. Verstraete (2009). Analysis and redesign of the production of measurement instru-
ments. Master’s thesis, University of Ghent.
J. H. Wilson & B. Keating (2002). Business Forecasting. McGraw-Hill Higher Educa-
tion.
SALES SURVEY 92
Appendix A
Sales survey
P&L STRUCTURE 93
Appendix B
P&L structure
P&L STRUCTURE 94
P&L STRUCTURE 95
OPTION ANALYSIS 96
Appendix C
Option analysis
Below you can find the scenarios table with the 3x3x3x3 combinations for the invest-
ment option.
Value Prob Cum Prob Growth Dollar Inflation Discount
0 0
1 3.076 0,0133 0,0133 5 0,90 3 0,00
2 2.353 0,0267 0,0400 5 0,90 3 0,50
3 1.629 0,0133 0,0533 5 0,90 3 1,00
4 3.244 0,0133 0,0667 5 0,90 4 0,00
5 2.479 0,0267 0,0933 5 0,90 4 0,50
6 1.714 0,0133 0,1067 5 0,90 4 1,00
7 3.410 0,0067 0,1133 5 0,90 5 0,00
8 2.601 0,0133 0,1267 5 0,90 5 0,50
9 1.792 0,0067 0,1333 5 0,90 5 1,00
10 4.981 0,0133 0,1467 5 1,48 3 0,00
11 4.258 0,0267 0,1733 5 1,48 3 0,50
12 3.535 0,0133 0,1867 5 1,48 3 1,00
13 5.287 0,0133 0,2000 5 1,48 4 0,00
14 4.522 0,0267 0,2267 5 1,48 4 0,50
15 3.757 0,0133 0,2400 5 1,48 4 1,00
16 5.601 0,0067 0,2467 5 1,48 5 0,00
17 4.791 0,0133 0,2600 5 1,48 5 0,50
18 3.982 0,0067 0,2667 5 1,48 5 1,00
19 5.212 0,0133 0,2800 5 1,60 3 0,00
20 4.489 0,0267 0,3067 5 1,60 3 0,50
21 3.766 0,0133 0,3200 5 1,60 3 1,00
22 5.535 0,0133 0,3333 5 1,60 4 0,00
OPTION ANALYSIS 97
23 4.770 0,0267 0,3600 5 1,60 4 0,50
24 4.004 0,0133 0,3733 5 1,60 4 1,00
25 5.866 0,0067 0,3800 5 1,60 5 0,00
26 5.057 0,0133 0,3933 5 1,60 5 0,50
27 4.247 0,0067 0,4000 5 1,60 5 1,00
28 10.538 0,0167 0,4167 10 0,90 3 0,00
29 8.902 0,0333 0,4500 10 0,90 3 0,50
30 7.265 0,0167 0,4667 10 0,90 3 1,00
31 11.031 0,0167 0,4833 10 0,90 4 0,00
32 9.298 0,0333 0,5167 10 0,90 4 0,50
33 7.566 0,0167 0,5333 10 0,90 4 1,00
34 11.526 0,0083 0,5417 10 0,90 5 0,00
35 9.692 0,0167 0,5583 10 0,90 5 0,50
36 7.858 0,0083 0,5667 10 0,90 5 1,00
37 14.783 0,0167 0,5833 10 1,48 3 0,00
38 13.146 0,0333 0,6167 10 1,48 3 0,50
39 11.510 0,0167 0,6333 10 1,48 3 1,00
40 15.588 0,0167 0,6500 10 1,48 4 0,00
41 13.855 0,0333 0,6833 10 1,48 4 0,50
42 12.122 0,0167 0,7000 10 1,48 4 1,00
43 16.417 0,0083 0,7083 10 1,48 5 0,00
44 14.583 0,0167 0,7250 10 1,48 5 0,50
45 12.749 0,0083 0,7333 10 1,48 5 1,00
46 15.297 0,0167 0,7500 10 1,60 3 0,00
47 13.661 0,0333 0,7833 10 1,60 3 0,50
48 12.025 0,0167 0,8000 10 1,60 3 1,00
49 16.140 0,0167 0,8167 10 1,60 4 0,00
50 14.407 0,0333 0,8500 10 1,60 4 0,50
51 12.675 0,0167 0,8667 10 1,60 4 1,00
52 17.010 0,0083 0,8750 10 1,60 5 0,00
53 15.175 0,0167 0,8917 10 1,60 5 0,50
54 13.341 0,0083 0,9000 10 1,60 5 1,00
55 20.211 0,0033 0,9033 15 0,90 3 0,00
56 17.426 0,0067 0,9100 15 0,90 3 0,50
57 14.642 0,0033 0,9133 15 0,90 3 1,00
58 21.135 0,0033 0,9167 15 0,90 4 0,00
59 18.184 0,0067 0,9233 15 0,90 4 0,50
60 15.233 0,0033 0,9267 15 0,90 4 1,00
61 22.069 0,0017 0,9283 15 0,90 5 0,00
OPTION ANALYSIS 98
62 18.943 0,0033 0,9317 15 0,90 5 0,50
63 15.816 0,0017 0,9333 15 0,90 5 1,00
64 27.327 0,0033 0,9367 15 1,48 3 0,00
65 24.542 0,0067 0,9433 15 1,48 3 0,50
66 21.757 0,0033 0,9467 15 1,48 3 1,00
67 28.782 0,0033 0,9500 15 1,48 4 0,00
68 25.831 0,0067 0,9567 15 1,48 4 0,50
69 22.880 0,0033 0,9600 15 1,48 4 1,00
70 30.285 0,0017 0,9617 15 1,48 5 0,00
71 27.158 0,0033 0,9650 15 1,48 5 0,50
72 24.032 0,0017 0,9667 15 1,48 5 1,00
73 28.189 0,0033 0,9700 15 1,60 3 0,00
74 25.404 0,0067 0,9767 15 1,60 3 0,50
75 22.620 0,0033 0,9800 15 1,60 3 1,00
76 29.709 0,0033 0,9833 15 1,60 4 0,00
77 26.758 0,0067 0,9900 15 1,60 4 0,50
78 23.807 0,0033 0,9933 15 1,60 4 1,00
79 31.280 0,0017 0,9950 15 1,60 5 0,00
80 28.154 0,0033 0,9983 15 1,60 5 0,50
81 25.027 0,0017 1,0000 15 1,60 5 1,00
100%
Below you can find the scenarios table with the 3x3x3x3 combinations for the shift
work option.
Value Prob Cum Prob Growth Dollar Inflation Discount
0 0
1 4.064 0,0133 0,0133 5 0,90 3 0,00
2 3.340 0,0267 0,0400 5 0,90 3 0,50
3 2.617 0,0133 0,0533 5 0,90 3 1,00
4 4.298 0,0133 0,0667 5 0,90 4 0,00
5 3.532 0,0267 0,0933 5 0,90 4 0,50
6 2.767 0,0133 0,1067 5 0,90 4 1,00
7 4.533 0,0067 0,1133 5 0,90 5 0,00
8 3.724 0,0133 0,1267 5 0,90 5 0,50
9 2.914 0,0067 0,1333 5 0,90 5 1,00
10 5.969 0,0133 0,1467 5 1,48 3 0,00
11 5.246 0,0267 0,1733 5 1,48 3 0,50
12 4.522 0,0133 0,1867 5 1,48 3 1,00
OPTION ANALYSIS 99
13 6.341 0,0133 0,2000 5 1,48 4 0,00
14 5.575 0,0267 0,2267 5 1,48 4 0,50
15 4.810 0,0133 0,2400 5 1,48 4 1,00
16 6.724 0,0067 0,2467 5 1,48 5 0,00
17 5.914 0,0133 0,2600 5 1,48 5 0,50
18 5.105 0,0067 0,2667 5 1,48 5 1,00
19 6.200 0,0133 0,2800 5 1,60 3 0,00
20 5.477 0,0267 0,3067 5 1,60 3 0,50
21 4.753 0,0133 0,3200 5 1,60 3 1,00
22 6.588 0,0133 0,3333 5 1,60 4 0,00
23 5.823 0,0267 0,3600 5 1,60 4 0,50
24 5.058 0,0133 0,3733 5 1,60 4 1,00
25 6.989 0,0067 0,3800 5 1,60 5 0,00
26 6.180 0,0133 0,3933 5 1,60 5 0,50
27 5.370 0,0067 0,4000 5 1,60 5 1,00
28 7.742 0,0167 0,4167 10 0,90 3 0,00
29 6.689 0,0333 0,4500 10 0,90 3 0,50
30 5.636 0,0167 0,4667 10 0,90 3 1,00
31 8.061 0,0167 0,4833 10 0,90 4 0,00
32 6.952 0,0333 0,5167 10 0,90 4 0,50
33 5.844 0,0167 0,5333 10 0,90 4 1,00
34 8.380 0,0083 0,5417 10 0,90 5 0,00
35 7.214 0,0167 0,5583 10 0,90 5 0,50
36 6.047 0,0083 0,5667 10 0,90 5 1,00
37 10.744 0,0167 0,5833 10 1,48 3 0,00
38 9.691 0,0333 0,6167 10 1,48 3 0,50
39 8.638 0,0167 0,6333 10 1,48 3 1,00
40 11.263 0,0167 0,6500 10 1,48 4 0,00
41 10.155 0,0333 0,6833 10 1,48 4 0,50
42 9.046 0,0167 0,7000 10 1,48 4 1,00
43 11.795 0,0083 0,7083 10 1,48 5 0,00
44 10.629 0,0167 0,7250 10 1,48 5 0,50
45 9.462 0,0083 0,7333 10 1,48 5 1,00
46 11.108 0,0167 0,7500 10 1,60 3 0,00
47 10.055 0,0333 0,7833 10 1,60 3 0,50
48 9.001 0,0167 0,8000 10 1,60 3 1,00
49 11.651 0,0167 0,8167 10 1,60 4 0,00
50 10.543 0,0333 0,8500 10 1,60 4 0,50
51 9.434 0,0167 0,8667 10 1,60 4 1,00
OPTION ANALYSIS 100
52 12.209 0,0083 0,8750 10 1,60 5 0,00
53 11.043 0,0167 0,8917 10 1,60 5 0,50
54 9.876 0,0083 0,9000 10 1,60 5 1,00
55 11.108 0,0033 0,9033 15 0,90 3 0,00
56 9.836 0,0067 0,9100 15 0,90 3 0,50
57 8.564 0,0033 0,9133 15 0,90 3 1,00
58 11.492 0,0033 0,9167 15 0,90 4 0,00
59 10.156 0,0067 0,9233 15 0,90 4 0,50
60 8.820 0,0033 0,9267 15 0,90 4 1,00
61 11.878 0,0017 0,9283 15 0,90 5 0,00
62 10.475 0,0033 0,9317 15 0,90 5 0,50
63 9.071 0,0017 0,9333 15 0,90 5 1,00
64 14.871 0,0033 0,9367 15 1,48 3 0,00
65 13.599 0,0067 0,9433 15 1,48 3 0,50
66 12.327 0,0033 0,9467 15 1,48 3 1,00
67 15.497 0,0033 0,9500 15 1,48 4 0,00
68 14.161 0,0067 0,9567 15 1,48 4 0,50
69 12.824 0,0033 0,9600 15 1,48 4 1,00
70 16.139 0,0017 0,9617 15 1,48 5 0,00
71 14.736 0,0033 0,9650 15 1,48 5 0,50
72 13.332 0,0017 0,9667 15 1,48 5 1,00
73 15.327 0,0033 0,9700 15 1,60 3 0,00
74 14.055 0,0067 0,9767 15 1,60 3 0,50
75 12.783 0,0033 0,9800 15 1,60 3 1,00
76 15.982 0,0033 0,9833 15 1,60 4 0,00
77 14.646 0,0067 0,9900 15 1,60 4 0,50
78 13.310 0,0033 0,9933 15 1,60 4 1,00
79 16.656 0,0017 0,9950 15 1,60 5 0,00
80 15.252 0,0033 0,9983 15 1,60 5 0,50
81 13.849 0,0017 1,0000 15 1,60 5 1,00
100,00%
OPTION ANALYSIS 101
Table C.1: Cash flows for investment option in x e 1000
2010 2011 2012 2013 2014 2015 2016 2017 2018
Yearly costs 1 2 3 4 5 6 7 8 9
Capital Investment 1.000 96 22 27 627 27 43 48 48
Total Goods 0 784 1.713 2.808 4.094 5.602 7.362 9.415 11.803Total Cost of Sales 0 460 701 1.058 1.406 1.840 2.330 2.987 3.687Total Operating 0 375 796 1.268 1.796 2.383 3.038 3.765 4.571ExpensesTotal Other 0 130 281 458 663 900 1.175 1.491 1.856Expenses
Total Cash out 0 1.749 3.491 5.592 7.958 10.726 13.905 17.658 21.918
Total Net Sales 0 2.730 5.921 9.639 13.956 18.958 24.739 31.409 39.090
Depreciation Tax -14 -17 -17 -18 -27 -28 -29 -30 -32Reduction
Cash flow -986 627 1.746 2.905 3.719 5.929 7.787 9.883 12.348after taxDiscount rate 18 %
Discounted -836 450 1.062 1.498 1.626 2.196 2.444 2.629 2.784
NPV cash flows 13.855after tax
Table C.2: Cash flows for shift work option in x e 1000
2011 2012 2013 2014 2015 2016 2017 2018
Yearly costs 2 3 4 5 6 7 8 9
Capital Investment 600
Total Goods 784 1.713 2.808 4.094 4.258 4.429 4.606 4.790Total Cost of Sales 332 591 1.002 1.449 1.506 1.565 1.626 1.689Total Operating Exp. 375 796 1.268 1.796 1.867 1.942 2.020 2.101Total Other Expenses 130 281 458 663 684 707 730 753
Total Cash out 1.621 3.382 5.535 8.002 8.316 8.642 8.981 9.333
Total Net Sales 2.730 5.921 9.639 13.956 14.411 14.881 15.365 15.864
Depr. Tax Reduction -8 -8 -8 -8 -8
Cash flow after tax 799 1.829 2.954 3.695 4.397 4.501 4.605 4.711Discount rate 18 %
Discounted 574 1.113 1.524 1.615 1.629 1.413 1.225 1.062
NPV cash flows 10.155after tax
OPTION ANALYSIS 102
Table C.3: Sensitivity analysis investment option (NPV values in x e 1000)
Description Base Low High Difference H-L
Sales growth 10% 5% 15%NPV values 13.855 4.522 25.831 21.309
Dollar value 1,48 0,9 1,6NPV values 13.855 9.298 14.407 5.109
Inflation percentage 4% 3% 5%NPV values 13.855 13.146 14.583 1.436
Cost production hall e 1.000.000 e 1.000.000 e 1.500.000NPV values 13.855 13.855 13.386 469
Cost administrative building e 600.000 e 600.000 e 900.000NPV values 13.855 13.855 13.731 124
Payroll cost increase 4% 2% 4%NPV values 13.855 14.363 13.855 508
Yearly discount increase 0,50% 0% 1%NPV values 13.855 15.588 12.122 3.465
Initial inefficiency decrease 10% 5% 15%NPV values 13.855 13.814 13.959 145
Yearly inefficiency decrease 1% 0% 2%NPV values 13.855 13.814 13.899 85
Price subcontracting hour e 40 e 35 e 50NPV values 13.855 14.108 13.349 760
OPTION ANALYSIS 103
Table C.4: Sensitivity analysis shift work option (NPV values in x e 1000)
Description Base Low High Difference H-L
Sales growth 10% 5% 15%NPV values 10.155 5.575 14.161 8.585
Dollar value 1,48 0,9 1,6NPV values 10.155 6.952 10.543 3.590
Inflation percentage 4% 3% 5%NPV values 10.155 9.691 10.629 938
Cost administrative building e 600000 e 600000 e 900000NPV values 10.155 10.155 10.030 125
Payroll cost increase 4% 2% 4%NPV values 10.155 10.623 10.155 468
Yearly discount increase 0,50% 0% 1%NPV values 10.155 11.263 9.046 2.217
Price subcontracting hour e 40 e 35 e 50NPV values 10.155 10.216 10.032 184
NEDERLANDSE SAMENVATTING 104
Appendix D
Nederlandse samenvatting
D.1 Inleiding
Magnetrol International N.V. (MINV) is een dochterbedrijf van het Amerikaanse Mag-
netrol International Incorporated. Het vervaardigt elektro-mechanische vlotter- en ver-
dringerniveauschakelaars en -meters, capacitieve niveauschakelaars, continue niveaumee-
tapparatuur met behulp van golfgeleide radar en vrijstralende radar enz. De afzetmarkt
situeert zich vooral in industrieen zoals de olie- en gas, petrochemische, energieproduc-
erende en chemische industrieen. Wereldwijd stelt Magnetrol meer dan 600 mensen te
werk.
De productieomgeving van MINV is geleidelijk aan over de jaren heen ontwikkeld,
vanaf zijn oprichting in 1971. 38 jaar later zit men op een punt waar de huidige
productievloer niet meer voldoet om antwoord te bieden aan de groei die Magnetrol
in de laatste jaren gekend heeft. Deze masterproef heeft als doel de Amerikaanse
eigenaar op een kwalitatieve en kwantitatieve manier te overtuigen van de noodzaak tot
capaciteitsuitbreiding. Door een profit and loss studie uit te voeren, kan er nagegaan
worden wat de invloed van investeringsbeslissingen zijn op de winst na belastingen,
alsook wordt er gekeken naar de NPV van de cashflows na belastingen.
Uiteraard zal een capaciteitsvergroting alleen niet volstaan om klaar te staan voor de
toekomst. Er is ook een flow en layout studie nodig om de efficientie van het systeem
te bewaken.
Een aantal problemen vallen onmiddellijk op wanneer we de productieomgeving beki-
jken. Hierna ziet u enkele indicatoren voor de nood aan extra capaciteit en ruimte:
� Vergrotende backlog : De geproduceerde waarde van vorig jaar bedroeg e 23,51
miljoen. MINV heeft in die periode voor e 26,1 miljoen verkocht. Het verschil
D.1 Inleiding 105
tussen beide bedragen kon dus niet geproduceerd worden, nl. e 2,59 miljoen. De
vraag van de markt is groter dan hetgeen geproduceerd kan worden, waardoor de
backlog aangroeit. De backlog is de waarde van de goederen die in behandeling
zijn door het systeem, zowel door administratie als productie. Zie het als een
wachtlijn waarbij de aankomstrate groter is dan de verwerkingsrate. Het effect
hiervan is nu al zichtbaar aan de langer wordende lead times, wat dan weer
negatief is vanuit een competitief standpunt.
� Meer uitbestedingsuren: Door de overbelasting in productie wordt er ook steeds
meer uitbesteed wat kan leiden tot kleinere winstmarges.
� Indicatoren op de productievloer : Het gebrek aan ruimte op de productievloer
brengt enerzijds extra material handling met zich mee en anderzijds is er ook een
verhoogd veiligheidsrisico. Zoals op foto’s 1.2 tot en met 1.6 te zien is, belemmert
WIP de transportgangen op de vloer. Door de trend van steeds langer wordende
probes is er plaatsgebrek aan de assembly tafels. Lange probes reiken bij hun
configuratie op de assembly tafels tot in de laskabines of spuitkabine. Los van
het capaciteitsprobleem kan de herschikking van de tafels hier al een oplossing
bieden.
� Kwaliteit van productie: In de huidige situatie zitten de ’vuile’ en ’propere’ op-
eraties in dezelfde ruimte waar zij in feite gesplitst zouden moeten zijn. De
reden hiervoor is dat de PCB’s vervuild kunnen worden door staaldeeltjes vanuit
machining. Anderzijds moet koolstofstaalproductie gescheiden zijn van roestvrij
staalproductie. Momenteel is er geen ruimte om deze splitsingen te maken.
De werkwijze van de thesis zit als volgt in elkaar:
� Gegevensanalyse van databank informatie: gegevens downloaden in een spread-
sheet en gespecialiseerde filters gebruiken zoals in figuur 2.2 om de informatie
overzichtelijk en werkbaar te maken.
� Op basis van deze gegevens de gemiddelde bewerkingstijden per product fam-
ilie en per departement berekenen, vooraf gegaan door het controleren van de
gegevens uit het eerste puntje.
� Een voorspelling maken van de toekomstige verkoop in eenheden van MINV en
deze gebruiken om de toekomstige product mix in te schatten. Deze mix wordt
dan geprojecteerd op een vooropgestelde groei van 5%, 10% en 15%.
D.2 Gegevensanalyse 106
� Deze voorspelling in combinatie met de directe productie uren per product familie
en per departement geeft een schatting van de benodigde productie uren over een
periode van 10 jaar in elk van de departementen.
� Vertaling van het aantal benodigde uren in VTE’s (voltijdse equivalenten) - reken-
ing houdend met inefficienties - en investeringen.
� Monte Carlo simulatie maken om de twee investeringen met elkaar te vergelijken:
– investering maken in nieuwe productiehal met bijhorend nieuw productiemid-
delen
– een 2-shiftensysteem invoeren
– niet uitbreiden wordt als referentie gebruikt om de cashflows te berekenen
die resulteren uit de investeringen
D.2 Gegevensanalyse
Zoals in de inleiding reeds vermeld werd, gaan we vooreerst de benodigde directe pro-
ductie uren per productfamilie per departement bepalen omdat ze niet rechtstreeks
beschikbaar zijn binnen de organisatie. In productie was men ter voorbereiding van
de implementatie van een nieuw ERP-systeem begonnen met het opmeten van pro-
ductietijden van een suborder in elk departement. Een suborder is een deel van een
bestelling bestaande uit een aantal identieke toestellen. Deze suborderboxen worden
gescand bij het starten en stoppen van de bewerking in een departement. Per dergeli-
jke start- en stoptijd wordt er een lijn in een data lijst bijgehouden. Een suborder kan
binnen een departement over meerdere lijnen beschikken. Dat gebeurt als er verschil-
lende bewerkingen moeten uitgevoerd worden of als de bewerking om een andere reden
gestopt dient te worden.
Door gebruik te maken van pivot tables in Excel is het mogelijk om deze ruwe informatie
op een gestructureerde manier per producttype weer te geven. Vervolgens worden de
resultaten gegroepeerd in 10 product families en worden de gemiddelde productietijden
per productfamilie bepaald. De koppeling tussen de productcodes en de productfamilies
gebeurt door gebruik te maken van de zogenaamde literature cross reference list MINV
(2009).
D.2 Gegevensanalyse 107
Table D.1: 10 Magnetrol productfamilies
Brand name Alternative name
Mechanicals Mechanical products
Modulevel Displacer transmitters
Eclipse Guided wave radar
Pulsar Radar
Gap Sensors Ultrasonic Contact
Thermatel Thermal dispersion
Kotron RF controls
Air Sonar Ultrasonic non-contact
Jupiter Magnetostrictive
Solitel Vibrating rod
Vooraleer we de bekomen data gebruiken, moet ze gecontroleerd worden. Hiertoe werd
er een meeting gepland met de productie manager van MINV om de data te corrigeren
en te fine tunen waar nodig. Hij maakte een voorstel van bepaalde afkappingsregels om
niet-correcte productie uren uit de analyse te filteren. Er zaten namelijk onrealistische
cijfers in de data lijst door bijvoorbeeld het vergeten uitscannen van suborders. Door
realistische afkappingsregels op te stellen werden foutieve data uit de gegevenslijst
verwijderd. Het resultaat van deze analyse wordt:
Table D.2: Directe productie uren per product familie per departement
Product familie Machine afdeling Lasafdeling Assemblage
Mechanical products 1,07 1,39 1,61
Displacer transmitters 1,67 2,43 2,19
Guided wave radar 1,73 1,67 1,49
Radar 0 0,31 0,96
Ultrasonic Contact 0,80 0,37 0,61
Thermal dispersion 1,42 0,33 1,19
RF controls 0 0,48 2,10
Ultrasonic non-contact 0 0,12 0,28
Magnetostrictive 0 0 0,47
Vibrating rod 0,08 0,02 0,29
Als een product in een bepaald departement weinig uren vergt, komt dit doordat deze
operatie hoofdzakelijk in de productie afdeling in Downers Grove (USA) gebeurt met
D.2 Gegevensanalyse 108
enkele kleine aanpassingen in Belgie.
Alvorens over te gaan naar de schatting van de toekomstige noden wordt er een korte
analyse gemaakt van de huidige productie situatie met betrekking tot eigen productie
en subcontracting. Er wordt bijvoorbeeld opgemerkt dat er een stijging is in de inef-
ficientie vanaf 2007 omdat er vanaf dat jaar veiligheidsvergaderingen moesten ingepland
worden, alsook door de overbelasting in productie met veel extra material handling als
gevolg.
De verkoopsvoorspellingen van de producten is het belangrijkste gedeelte van de gegevens-
analyse. De basis voor deze analyse is een lijst met historische verkoopscijfers van
MINV op jaarbasis gegeven. Eclipse is een product dat vanaf 1998 in productie werd
genomen, waardoor er maar 11 elementen in de data serie zitten. Deze beperkte
gegevensbeschikbaarheid zal een grotere onzekerheid met zich meebrengen omtrent de
voorspellingen, zeker omdat er 10 jaar in de toekomst geschat dient te worden. Hoe
verder men in de toekomst probeert te voorspellen, hoe groter de onzekerheid. Daarom
ook werd de kwantitatieve forecast gestaafd met een kwalitatief marktonderzoek bij de
verkopers. De resultaten van de sales survey zullen ons een goed idee geven van de
toekomstige product mix van MINV. Het verkoopskanaal is immers een rijke bron van
informatie omdat zij dag in dag uit in contact komen met klanten en ze zijn het dichtste
contact dat een bedrijf met zijn klanten heeft. De resultaten van het marktonderzoek
zijn te zien in tabel 2.4. De cijfers 1 t.e.m. 5 stellen respectievelijk sterk dalende en
sterk stijgende verkoop voor.
De kwantitatieve voorspelling wordt uitgevoerd met behulp van de ForecastX software,
geleverd bij het boek Wilson & Keating (2002). Voor elk product wordt de tijdserie
ingegeven en wordt er gekozen op basis van welke foutmetingen het voorspellingsmodel
geselecteerd moet worden. Zo kan men kiezen voor bijvoorbeeld MSE, SSE en RMSE.
Afhankelijk van de gekozen meting zal het programma het model kiezen waarvoor de
meting minimaal is. De gekozen modellen vindt u in tabel 2.6. Door de jonge leeftijd
van sommige product families waren er weinig historische gegevens om handen, wat
de voorspelling dan enigszins bemoeilijkte. De resultaten van de verkoopsvoorspelling
worden in alle geval gecontroleerd aan de hand van de resultaten van een marktonder-
zoek bij de verkoopsmensen van MINV. Vooral bij gebrek aan historische gegevens van
jonge producten hebben we ons hierop moeten baseren.
De resultaten van de verkoopsvoorspelling worden gebruikt om voor de komende jaren
een geschatte product mix op te stellen. De product mix van 2009 en 2018 ziet u
respectievelijk in figuur 2.19 en figuur 2.20 in sectie 2.4.4. De getallen achter de pro-
ductnamen zijn in % weergegeven, om op te tellen tot 100%. Deze product mix wordt
D.2 Gegevensanalyse 109
geprojecteerd op de voorafbepaalde jaarlijkse groei van 10%, met als worst case sce-
nario een groei van 5% en 15% groei als best case scenario. Merk op dat Eclipse de
sterkste groeier is en 41% van de geproduceerde eenheden in 2018 zou uitmaken, wat
eventueel justifieert om een aparte Eclipse area te voorzien, een product layout als u wil,
waar alle processen die benodigd zijn om het product te maken in lijn zijn opgesteld.
Nu we de benodigde directe productie uren per product familie en per departement
hebben, gecombineerd met een voorspelling van de toekomstige verkoop, kunnen we
overgaan tot de berekening van de toekomstige nood aan productie capaciteit, uitge-
drukt in VTE’s. Om te beginnen projecteren we de product mix op verschillende
groeipercentages zoals eerder gezegd. In verdere tekst wordt er steeds gewerkt met
een groei van 10%, zijnde de jaarlijkse groeitarget van Magnetrol. De twee overige
scenario’s worden verwerkt in de Excel assessment tool die wordt bijgevoegd bij deze
masterproef. De scenario’s worden ook verwerkt in de investeringsanalyse van sectie
4. De projectie levert de toekomstig benodigde aantallen en, in combinatie met de
benodigde productieuren, de nodige directe uren per product voor de komende jaren
op. Bij de berekening van de VTE’s, en dus de incorporatie van indirecte uren, werd
er rekening gehouden met eventueel een verbeterde efficientie. Er kan met een extra
vloeroppervlakte immers gewerkt worden aan de reductie van indirecte uren door over-
bodig transport of blokkeren van transport te vermijden. Ook zal de material handling
binnen assemblage vlotter kunnen verlopen. Om het voordeel van verbeterde efficientie
aan te tonen wordt er berekend hoeveel manjaren uitgespaard kunnen worden en welk
bedrag hiermee overeenstemt. De resultaten van deze analyse worden samengevat in
figuren 2.28 tot en met 2.30.
Bij een stijgende verkoop heeft Magnetrol voorzien om de ondersteunende manpower
met volgende percentages te verhogen:
� Als de verkoop jaarlijks met 5% stijgt, dan mag de headcount van de onderste-
unende functies jaarlijks stijgen met 2,66%.
� Als de verkoop jaarlijks met 10% stijgt, dan mag de headcount van de onderste-
unende functies jaarlijks stijgen met 4,14%.
� Als de verkoop jaarlijks met 15% stijgt, dan mag de headcount van de onderste-
unende functies jaarlijks stijgen met 5,45%.
Een andere opmerking in verband met schaalbaarheid is dat wanneer de benodigde
directe uren verdubbelen (verkoop verdubbelt), dan zullen in het geconstrueerde model
de totale indirecte uren met dezelfde factor stijgen. Dat wil in feite zeggen dat ook de
D.3 Financiele evaluatie 110
supervisie met die factor vermenigvuldigd wordt, wat in praktijk natuurlijk niet perse
opgaat. Het kan zijn dat Magnetrol bijvoorbeeld de helft meer laskabines kan zetten
en lassers kan tewerkstellen zonder dat het daarvoor de supervisie door een leadman
moet verhogen.
D.3 Financiele evaluatie
D.3.1 Inleiding
Na de berekening van de benodigde VTE’s voor de komende jaren kan er overgegaan
worden tot de financiele evaluatie. We maken gebruik van de profit & loss statement
van Magnetrol International N.V. om drie gevallen te bekijken:
� Er wordt beslist om niet verder te investeren in de Belgische productiefaciliteiten
van Magnetrol International.
� Een investering wordt gemaakt in:
– Het verdubbelen van de bestaande productievloer met herschikking van
het bestaande productie apparaat en het installeren van extra productie-
uitrusting.
– Uiteraard moet verdere verkoops- en productiegroei ondersteund worden
door het uitbreiden van de dienstenactiviteiten. Het bestaande adminis-
tratieve gebouw heeft echter weinig ruimte om extra personeel te plaatsen,
zodus moet er overwogen worden om het bestaande gebouw uit te breiden.
Deze uitbreiding zou pas in gebruik genomen worden vanaf 2014.
� Vanuit het moederbedrijf MII wordt er gevraagd om ook een 2-shiftensysteem te
overwegen.
Alvorens de berekeningen aan te vatten wordt er een analyse gemaakt van de P&L-
structuur van MINV. Alle rubrieken worden bekeken en er wordt een onderscheid
gemaakt tussen deterministische en stochastische rubrieken. Stochastische rubrieken
zijn rubrieken waarvan de waarde afhangt van input variabelen die zich moeilijk laten
voorspellen en waarvan de waarde dus op voorhand niet geweten is. Deterministische
rubrieken worden vaak berekend als som van andere, dikwijls stochastische rubrieken.
We gaan na welke rubrieken beınvloed worden door beslissingen als extra blue col-
lars, white collars, productie-uitrusting, nieuwe productiehal en een nieuw bureauge-
bouw. Zoals eerder vermeld werd, zullen de toekomstmogelijkheden in drie gevallen
onderverdeeld worden.
D.3 Financiele evaluatie 111
Op het einde van de ’investerings-’ en ’shift werk’ case wordt er een cumulatieve dis-
tributie opgesteld. Hierbij worden er voor elk van de meest onzekere beslissingen van
een bepaalde case enkele waarden meegegeven en worden de verschillende mogelijke
winsten berekend en uitgezet in een grafiek. Een dergelijke grafiek laat zich bijvoor-
beeld lezen als volgt: er is 85% kans dat onze winst meer dan e 5 miljoen zal bedragen.
De P&L analyse wordt voorafgegaan door een keuze aan parameters die we in de
assessment tool zullen incorporeren. Zo zijn er:
� economische parameters als de aangroei van lijstprijzen, de dollar/euro koers, het
duurder worden van materialen en de sales target van MINV;
� P&L relaties die de verhouding uitdrukken van een bepaalde P&L rubriek op
een grote. Deze relaties worden gebruikt omdat bepaalde rubrieken zich heel
moeilijk vooraf laten bepalen en het aanvaardbaar is om deze als percentage van
een andere, welgekozen rubriek te nemen. Zo zal de verhouding genomen worden
van Other manufacturing material op Total Other Sales ;
� parameters met betrekking tot arbeid zoals de toename van dienstenheadcount
(afhankelijk van de groei in sales), de aangroei van de lonen (in feite ook economis-
che parameter, bevat loonsopslag alsook indexering), een toeslag voor shift werk,
gemiddelde loonskost in productie ...;
� uitbestedingsparameters zoals de uitbestedingskost per machine-uur en prijssti-
jging van uitbestedingen;
� investeringsparameters zoals de duur van afschrijvingsperioden, aantal laskabines,
kost van een laskabine en afzuiginstallatie, kost van een nieuwe productiehal...
Na de individuele analyses komt er een vergelijk tussen de verschillende cases alsook
een analyse van de invloed van de dollar/euro koers op de winst van MINV. Deze topic
vormt al jaren een aandachtspunt en er werd verwacht deze invloed te kwantificeren
aan de hand van de assessment tool.
D.3.2 Resultaten
Huidige productie situatie behouden
Als er niet wordt geınvesteerd in nieuwe productiecapaciteit en we weten dat de huidige
situatie verzadigd is, zou er enkel meer geproduceerd kunnen worden mits uit te best-
eden. Hier zitten we echter met de beperking van de capaciteit van de assemblage die
D.3 Financiele evaluatie 112
sowieso in house gebeurt. Assemblage blijft de bottleneck in het productieproces zodat
uitbreiding van de verkoop vrijwel onmogelijk is zonder een explosie van de backlog.
Dat heeft lange lead times tot gevolg wat de klantenservice sterk vermindert.
Deze situatie wordt gedetailleerd uitgewerkt in sectie 3.4.1. De resultaten van de
analyse zijn voorgesteld in grafiek D.1.
Figure D.1: Financiele evaluatie in geval van geen bijkomende investeringen
Investeringsoptie
De investeringen zouden ingevoerd worden in 2 fasen om de investeringslast minder
zwaar te laten doorwegen op de balans.
� Fase 1
– Bouwen van nieuwe productiehal en in gebruik nemen tegen 2011.
– Nieuwe blue collars aannemen volgens de berekende VTE aantallen of vol-
gens de noden. Ook nieuwe white collars om de toegenomen verkoop te
ondersteunen, vooral na 2015.
– Nieuwe productie uitrusting ter uitbreiding van de bestaande capaciteit vol-
gens het aantal VTE’s (machining wordt verder aanvullend aan huidige ca-
paciteit uitbesteed) met herschikking van de bestaande layout.
� Fase 2
– Nieuwe kantoren worden in 2015 in gebruik genomen. In dit nieuw kan-
toorgebouw zal er ook een nieuwe demonstratieruimte en ruimere vergaderzaal
voorzien worden.
D.3 Financiele evaluatie 113
Er wordt ook gewezen op de noodzaak om projecten te organiseren om tussenstocks
te verminderen en dit om optimaal van de verkregen ruimte gebruik te kunnen maken.
Een verandering dringt zich op in de manier van productie inplannen en de processen
moeten op elkaar afgestemd worden in de mate van het mogelijke.
De resultaten van deze P&L analyse worden vertaald in grafiek D.2.
Figure D.2: Financiele evaluatie in geval van investeringen
2-shiftensysteem
Ook hier wordt dezelfde P&L analyse toegepast met de relevante parameters van hier-
voor. Bij het implementeren van een 2-shiftensysteem wordt een surplus van 15% op
de loonkost berekend om de kosten i.v.m. shiftwerk te dekken. De resultaten worden
gegeven in de grafiek van figuur D.3.
D.3 Financiele evaluatie 114
Figure D.3: Financiele evaluatie in geval van een 2-shiftensysteem
Dollar analyse
De dollar/euro koers kende een grote evolutie sinds het begin van de 21e eeuw. Voor
midden 2002 kreeg men minder dan 1 dollar per euro, momenteel bedraagt dat 1,3
dollar per euro. De laatste maanden kende de dollarkoers een verloop van 1,6 naar
1,3 dollar per euro. Gezien de grote hoeveelheid goederen die MINV aankoopt bij
het moederbedrijf in Amerika heeft deze koers een aanzienlijke invloed op de balans
van MINV, welke niet gecompenseerd wordt door het moederbedrijf. Naarmate MINV
meer goederen aankoopt, wat het geval is voor de shiftwerk- en investeringscase, zal
deze invloed nog groter zijn. De resultaten van de analyse worden in drie grafieken
uitgezet, een voor elke case. Hierbij wordt er gekeken naar 4 koersen $/e 0,9 - 1,28 -
1,48 - 1,6. De koers van 1,28 $/e is de koers op het ogenblik van schrijven, 1,48 $/e is
de gemiddelde gewogen koers voor MINV uit 2008. De resultaten kan de lezer vinden
in figuren 3.12 tot en met 3.14.
De dollar koers is duidelijk een factor die niet mag genegeerd worden. In elk van de
drie gevallen heeft een verschuiving van de lange termijn dollar koers een grote impact
op de winst van MINV. Als de lezer even een kijkje neemt in de P&L tabellen uit de
assessment tool is het relatieve belang van de goederen uit de USA, Goods MII, al snel
duidelijk.
D.4 Vergelijking van de cases 115
D.4 Vergelijking van de cases
D.4.1 Monte Carlo simulatie
De vergelijking tussen de resultaten van de 3 cases worden voor u samengevat in figuur
D.4. In onderstaande figuur worden alle input variabelen op hun basisch dwaarde
ingesteld en wordt er gewerkt met de PAT uit de P&L tabellen (geen NPV).
Figure D.4: PAT voor een periode van 10 jaar en voor de 3 cases
Het bovenstaande volstaat echter niet om een volwaardige conclusie te trekken van
welke case beter is, 2-shift systeem of investeringscase. Er wordt namelijk geen rekening
gehouden met de tijdswaarde van geld, noch met de absolute investeringsbedragen. Met
dat laatste wordt bedoeld dat in de bovenstaande P&L analyse enkel gekeken wordt
naar de afschrijvingen die resulteren uit de investeringen, terwijl de reele invloed in
feite het investeringsbedrag op een zeker tijdstip is. De tijdswaarde van e 1.000.000
geınvesteerd in 2010 is groter dan de som van gelijke afschrijvingen over 20 jaar van de
investering. Langs de andere kant wordt er geen rekening gehouden met de variantie
die kan zitten op de input variabelen.
We gaan in plaats hiervan gebruik maken van een Monte Carlo simulatie waarbij we
de NPV van de cashflows na belastingen gaan gebruiken als beslissingscriterium. De
input variabelen krijgen hier in plaats van enkel een gemiddelde waarde ook nog twee
uiterste waarden mee om het effect van variantie in rekening te brengen.
D.4 Vergelijking van de cases 116
De werkwijze voor dergelijke analyse zit als volgt in elkaar:
� De variabelen die het grootste invloed hebben op de NPV van de cashflows na
belasting selecteren via een gevoeligheidsanalyse (zie tabel C.3 and C.4).
� Gebruik de vier geselecteerde variabelen met hun 3 waarden elk. Stel een tabel
op met alle mogelijke combinaties om vervolgens de probabiliteit van elk sce-
nario te berekenen. De scenario-probabiliteiten worden berekend door de deel-
probabiliteiten uit tabel 4.3.2 van elke variabele met elkaar vermenigvuldigd.
Neem nu bijvoorbeeld het scenario dat er 5% verkoopsaangroei is, dat 1 euro
1,48 dollar waard is, er 4% inflatie is en dat de kortingen jaarlijks met 1% ver-
hoogd worden. Dit scenario heeft dan een probabiliteit om op te treden van
0, 45.0, 33.0, 4.0, 25 = 0, 0150 of 1,5 % kans. Naast een kolom met individuele
probabiliteiten van de scenario’s wordt er ook een kolom met de cumulatieve
probabiliteit voorzien.
� Vervolgens worden er door Excel 10.000 random getallen (tussen 0 en 1) gegenereerd
voor de twee opties. Elk van die random getallen R wordt gebruikt om een sce-
nario te selecteren. Dat gebeurt door in Excel een zoekopdracht te vervullen
waarbij er in de kolom met cumulatieve probabiliteiten gezocht wordt naar R.
Gezien Excel steeds de ondergrens selecteert van het interval waar R inligt, laten
we als output de NPV van de volgende rij weergeven. Scenario’s met een grotere
probabiliteit hebben zo meer kans om voor te komen in de random reeks. Ver-
volgens wordt er met de 10.000 gegenereerde NPV’s een cumulatieve distributie
opgesteld zowel voor het 2-shiftensysteem als voor de investeringscase. De verkre-
gen grafiek D.4.2 laat toe een beter vergelijk te maken tussen beide opties.
D.4.2 Resultaten
De resultaten van de simulatie vindt u in onderstaande grafiek.
D.4 Vergelijking van de cases 117
Figure D.5: Cumulatieve probabiliteiten uit Monte Carlo simulatie
Rekening houdend met de cumulatieve distributies kan er besloten worden dat in 40%
van de gevallen het shiftensysteem beter is dan de investeringsoptie (de rode lijn ligt
rechts van de blauwe lijn). In 60% van de gevallen is de investeringsoptie beter. De
horizontale afstand tussen de grafieken geeft het verschil in NPV weer. Natuurlijk kan
deze analyse niet voorstellen wat er in werkelijkheid zal gebeuren, maar het geeft een
goed idee van het verschil tussen het shiftensysteem en de investeringsoptie.
In 30% van de gevallen is de NPV van de cashflows afkomstig uit de investeringsoptie
beneden e 5.000.000 en in minder dan 20 % van de gevallen is de NPV van de cashflows
uit het shiftensysteem beneden e 5.000.000. In 10 % van de gevallen is de NPV van de
shiftwerkoptie boven e 11.000.000, terwijl dat bij de investeringsoptie tot 45 % gaat.
We laten de simulatie van 10000 random gevallen 200 keer uitvoeren en van elk van de
simulaties wordt het gemiddelde bijgehouden. Het gemiddelde van die 200 gemiddeldes
berekenen geeft dan e 10.093.171 voor de investeringsoptie en e 7.824.048 voor de
shiftoptie, in het voordeel van de investeringsoptie.
Als we ons baseren zowel op de cumulatieve distributie als op de gemiddeldes, blijkt de
investeringsoptie in beide gevallen de betere oplossing. Voor de lage waarden van NPV
in de linkerstaart van de distributie blijkt de shiftoptie beter te zijn. Zoals verwacht,
gebeurt dat in geval van een lage verkoopsaangroei (5 %). Een verklaring hiervoor is te
vinden in het feit dat bij lage verkoopsaangroei de productie in het shiftensysteem even
lang kan blijven aangroeien als in de investeringsoptie zonder gesatureerd te worden (in
D.4 Vergelijking van de cases 118
onze studieperiode van 10 jaar). In de investeringscase zijn er echter grote investeringen
gemaakt om de verkoopsgroei bij te benen, investeringen die niet gemaakt werden in
de shiftcase en zwaar doorwegen op de NPV van de investeringsoptie.
Naast de financiele vergelijking moet er ook gekeken worden naar de sociale impact
van shiftwerk. Eerst en vooral moet het management de huidige werkkrachten laten
instemmen om in shiften te werken en daarnaast moet het nog extra werkkrachten
vinden die bereid zijn om in shift te werken. In het industriepark van Zele zijn er slechts
weinige bedrijven die in shiften werken wat de zoektocht in deze area nog bemoeilijkt.
LIST OF FIGURES 119
List of Figures
1.1 US Dollars per barrel, OK WTI spot price FOB . . . . . . . . . . . . . 3
1.2 Shop floor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.3 Shop floor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.4 Shop floor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.5 Shop floor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.6 Shop floor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.7 Manoeuvring long probe . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.8 Probe blocking safe passage . . . . . . . . . . . . . . . . . . . . . . . . 9
1.9 Inefficiency percentages in production . . . . . . . . . . . . . . . . . . . 13
2.1 Raw data list . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.2 Pivot table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.3 Total production hours per department . . . . . . . . . . . . . . . . . . 19
2.4 Direct production hours per department . . . . . . . . . . . . . . . . . 20
2.5 Direct production hours per department . . . . . . . . . . . . . . . . . 20
2.6 Forecast illustration dr. ir. R. Van Landeghem (2008) . . . . . . . . . . 21
2.7 Forecast Eclipse Graph . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.8 Forecast sales units Air Sonar . . . . . . . . . . . . . . . . . . . . . . . 29
2.9 Forecast sales units Eclipse . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.10 Forecast sales units Gap Sensors . . . . . . . . . . . . . . . . . . . . . . 30
2.11 Forecast sales units Jupiter . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.12 Forecast sales units Kotron . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.13 Forecast sales units Mechanicals . . . . . . . . . . . . . . . . . . . . . . 32
2.14 Forecast sales units Modulevel . . . . . . . . . . . . . . . . . . . . . . . 32
2.15 Forecast sales units Pulsar . . . . . . . . . . . . . . . . . . . . . . . . . 33
2.16 Forecast sales units Solitel . . . . . . . . . . . . . . . . . . . . . . . . . 34
2.17 Forecast sales units Thermatel Switch . . . . . . . . . . . . . . . . . . . 34
2.18 Forecast sales units Thermatel Transmitter . . . . . . . . . . . . . . . . 35
2.19 Pie chart of the forecasted product mix for 2009 . . . . . . . . . . . . . 36
2.20 Pie chart of the forecasted product mix for 2018 . . . . . . . . . . . . . 36
LIST OF FIGURES 120
2.21 Forecast sales of largest product families . . . . . . . . . . . . . . . . . 40
2.22 Total direct hours in the machining department . . . . . . . . . . . . . 40
2.23 Total direct hours in the welding department . . . . . . . . . . . . . . . 41
2.24 Total direct hours in the asembly department . . . . . . . . . . . . . . 41
2.25 Inefficiency reduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
2.26 Influence of inefficiency reduction on FTEs . . . . . . . . . . . . . . . . 43
2.27 Difference between 1% and 5% annual inefficiency reduction . . . . . . 44
2.28 Full time equivalents . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
2.29 Full time equivalents with better efficiency . . . . . . . . . . . . . . . . 45
2.30 Cumulative reduction in mean years . . . . . . . . . . . . . . . . . . . . 46
2.31 Full time equivalents . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
2.32 Full time equivalents with better efficiency . . . . . . . . . . . . . . . . 47
2.33 Cumulative reduction in mean years . . . . . . . . . . . . . . . . . . . . 47
2.34 Full time equivalents . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
2.35 Full time equivalents with better efficiency . . . . . . . . . . . . . . . . 48
2.36 Cumulative reduction in mean years . . . . . . . . . . . . . . . . . . . . 49
3.1 The influence of discount on list prices of Magnetrol . . . . . . . . . . . 59
3.2 Total other sales divided by total invoiced . . . . . . . . . . . . . . . . 60
3.3 Financial evaluation of ’no investment’ case . . . . . . . . . . . . . . . 64
3.4 Labour costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
3.5 Financial evaluation of ’investment’ case if 10% growth occurs . . . . . 74
3.6 Headcount in the production departments . . . . . . . . . . . . . . . . 75
3.7 Corrected headcount in the production departments . . . . . . . . . . . 76
3.8 Operating expenses - payroll cost/total cost . . . . . . . . . . . . . . . . 77
3.9 Cost evolution of service departments . . . . . . . . . . . . . . . . . . . 78
3.10 Financial evaluation of the ’shift work’ case . . . . . . . . . . . . . . . 79
3.11 Dollar/Euro exchange rate from 2002 to 2009 BC investments (2009) . 80
3.12 Influence of dollar/euro exchange rate on the profit in the ’no investment’
case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
3.13 Influence of dollar/euro exchange rate on the profit in the ’investment’
case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
3.14 Influence of dollar/euro exchange rate on the profit in the ’shift work’ case 82
4.1 PAT over a period of 10 years for the 3 cases . . . . . . . . . . . . . . . 84
4.2 Darts game illustration of Monte Carlo simulation . . . . . . . . . . . . 88
4.3 Cumulative probabilities as perceived in Monte Carlo simulation . . . . 89
D.1 Financiele evaluatie in geval van geen bijkomende investeringen . . . . 112
D.2 Financiele evaluatie in geval van investeringen . . . . . . . . . . . . . . 113
LIST OF FIGURES 121
D.3 Financiele evaluatie in geval van een 2-shiftensysteem . . . . . . . . . . 114
D.4 PAT voor een periode van 10 jaar en voor de 3 cases . . . . . . . . . . 115
D.5 Cumulatieve probabiliteiten uit Monte Carlo simulatie . . . . . . . . . 117
LIST OF TABLES 122
List of Tables
2.1 10 Magnetrol product families . . . . . . . . . . . . . . . . . . . . . . . 15
2.2 Direct production hours per product family and per department . . . . 18
2.3 Sales units forecast Eclipse . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.4 Sales Survey: scores as defined in text . . . . . . . . . . . . . . . . . . . 24
2.5 Summary Sales Survey . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.6 Selected forecasting models . . . . . . . . . . . . . . . . . . . . . . . . . 28
2.7 Forecasted quantities 2009-2013 . . . . . . . . . . . . . . . . . . . . . . 37
2.8 Forecasted quantities 2014-2018 . . . . . . . . . . . . . . . . . . . . . . 38
2.9 Forecast Projection on a 10% target growth 2009-2013 . . . . . . . . . 39
2.10 Forecast Projection on a 10% target growth 2014-2018 . . . . . . . . . 39
3.1 Economic parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
3.2 Profit & Loss relations . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
3.3 Labour parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
3.4 Subcontracting parameters . . . . . . . . . . . . . . . . . . . . . . . . . 54
3.5 Investment parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
3.6 Costs welding station . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
3.7 Facility costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
4.1 Chosen variables for a sensitivity analysis . . . . . . . . . . . . . . . . . 86
4.2 Probabilities of variables’ values . . . . . . . . . . . . . . . . . . . . . . 87
C.1 Cash flows for investment option in x e 1000 . . . . . . . . . . . . . . . 101
C.2 Cash flows for shift work option in x e 1000 . . . . . . . . . . . . . . . 101
C.3 Sensitivity analysis investment option (NPV values in x e 1000) . . . . 102
C.4 Sensitivity analysis shift work option (NPV values in x e 1000) . . . . 103
D.1 10 Magnetrol productfamilies . . . . . . . . . . . . . . . . . . . . . . . 107
D.2 Directe productie uren per product familie per departement . . . . . . 107