BICG Buy vs Build
-
Upload
zafar-imam-khan -
Category
Documents
-
view
234 -
download
0
Transcript of BICG Buy vs Build
-
8/12/2019 BICG Buy vs Build
1/7
1.1 OverviewThe following project comparisons
are meant to serve as a guideline for or-
ganizations involved in the selection
process with Oracle, and who are consid-
ering purchasing the pre-built analytic
applications (which contain pre-written
ETL, data models, and metadata reposi-
tories) versus building custom datamartsor a data warehouse against operational
applications from Oracle including Ora-
cle E-Business Suite (EBS), PeopleSoft,
Siebel, JD Edwards, and similar systems.
BI Consulting Group has completed
more than 250 OBI EE projects, with
slightly more than half of those based
upon the pre-built applications from
Oracle. Its been our experience that the
pre-built analytic applications from Ora-
cle typically meet at least 70% - 80% of
EBS, PeopleSoft, Siebel, or JD Edwards
customers data needs, which results in a
dramatic reduction in the level of effort
required to deploy a DW and BI solution
on top of those systems. To put this in
real terms, our cost and duration esti-
mates for building custom datamarts
against Oracle EBS, Peoplesoft, Siebel, or
JD Edwards run at least 2.5 to 3.5 times
our estimates for deploying the pre-built
applications. And, thats for a team that
knows the data models of Oracle EBS,
PeopleSoft, Siebel, or JD Edwards inti-
mately the estimates would be signifi-
cantly higher (or the result significantly
lower) for a team not familiar with those
operational data models.
The most common question that
arises when organizations are investigat-
ing a deployment of the pre-built applica-
tions versus a custom developed data
warehouse, is "how can applications that
are 'generically' designed, work for my
organization"? That's actually an easyanswer, and follows the 80/20 rule. Afterbuilding and deploying more than 300
data warehouses, we've reached the con-
clusion that for a given subject area (for
example, Financial Analytics) 80% or
more of the average data warehouse isbased upon non-unique content. This
content is typically the same regardless of
whether your business creates pharma-
ceuticals, flies airplanes, or makes hard
drives.
Data elements such as Address-
Field1, ZipCode or AccountID
might have different names from business
to business, but those non-unique fields
take up the vast majority of development
time during the average data warehouse
project. Its the last 15% - 20% of devel-
opment that makes the unique differences
between industries, as well as the differ-
ences in business strategies from company
to company, even if those businesses sell
the same types of products to the same
types of customers.
BI CONSULTING GROUP WHITEPAPER VERSION 2.0 FY08
- 1-
1.1OVERVIEWThis whitepaper explores the
relative pros and cons of
deploying pre-built analytic
applications from Oracle versus
building a custom data
warehouse against Oracle
E-Business Suite (EBS),
PeopleSoft, Siebel, JD Edwards or
Similar Systems
1.2DEPLOYMENT OFPREBUILTAPPLICATIONSAGAINST EBS,PEOPLESOFT, SIEBEL, JDEDWARDS OR SIMILARSYSTEMSProject Schedule and Milestones
for Deployment of Pre-Built
Applications against EBS,
PeopleSoft, Siebel, or JD
Edwards
1.3DEPLOYMENT OFPREBUILTAPPLICATIONSAGAINST A NON-ORACLE BACKENDProject Schedule and Milestones
for Deployment of Pre-Built
Applications against a non-
Oracle backend
1.4CUSTOM WAREHOUSEBUILD AGAINST EBS,PEOPLESOFT, SIEBEL, JDEDWARDS OR SIMILARSYSTEMSProject Schedule and Milestones
for Custom Warehouse Build
against EBS, Peoplesoft, Siebel, JD
Edwards or Similar Systems
Oracles Pre-Built Analytic ApplicationsVersus Building a Custom Warehouse
0
500
1,000
290
925
17 weeks 34 Weeks
Buy Build
Aver
PersDays
Project Duration
Cost Estimates of deploying Custom Datamartstypically run 2.5 to 3.5 times the Cost Estimates
of deploying pre-built applications.
Copyright 2007, BI Consulting Group, LLC. All rights reserved. No part of this material may be reproduced, stored in a retrieval system or transmitted by anymeans, electronic, mechanical, photocopy, recording or otherwise, without written permission from the authors.
-
8/12/2019 BICG Buy vs Build
2/7
Why have Pre-Built Analytic
Applications Historically Failed?The Theory behind pre-built analytic applications has been
around for a number of years, but nearly all have failed to deliver.
Traditional BI vendors have attempted to either package pre-built
metadata, mapped directly to the operational tables of CRM,
ERP, HR and other systems (including Oracle E-Business Suite,Siebel, PeopleSoft, JD Edwards and SAP); or, theyve attempted to
write a small number of extracts to move that data into some
form of datamart or data warehouse. Historically, these attempts
have failed for two reasons. First, using the approach of mapping
directly to these operational systems failed for the simple fact that
theyve skipped one of the most important steps of business intel-
ligence, namely the data warehouse. Those BI vendors glossed
over that fact by using names like Siebel Adaptor or Oracle
EBS Connector, and conveniently ignored 20 years of data
warehouse best practices.
This style of pre-built application failed on two fronts. First,
they simply couldnt provide the performance required of a BIapplication against these operational systems (and also brought
many mission critical applications to their knees), and at best, they
mapped a minor subset of all possible metrics, KPIs and report-
able attributes. The second approach taken by these traditional
BI vendors had a glimmer of hope, since they attempted to follow
the more accepted approach to business intelligence, namely
building a datamart or data warehouse on top of these opera-
tional systems, using an industry accepted ETL tool, and the de-
sign and development of one or more star-schemas. While good
in theory, the only BI vendor who was able to deliver on this
promise, was the only BI vendor who also happened to be a CRM
vendor, namely Siebel Systems (and now Oracle). Reasons vary
why other BI vendors failed, but the greatest reason might also be
the simplest, which was a lack of investment. Building a true data
warehouse against a complex operational system takes time and
money. Siebel was able to make this investment, while other BI
vendors were not.
Where is the real value of
Pre-Built Analytic Applications?When it comes to pre-built analytic applications, all is not
equal when it comes to the value from each component. The
diagram to the upper right is a close approximation of the contri-
bution that each component typically has to individual projects.
For example, if the entire pre-built stack adds 100% value in a
completely vanilla environment against a known backend system
such as Oracle EBS, then only 20% - 30% must be made up if the
pre-built analytic stack is pointed to a backend system not sup-
ported by the out-of-the-box extracts. In real world terms, that
means that if Oracle Financial Analytics (which has extracts writ-
ten for Oracle E-Business Suite and PeopleSoft) was pointed in-
stead to a non-EBS or PeopleSoft backend, the additional work
required, might be no more than 20% - 30% added to the project.
Again, in real world terms, if a "typical" deployment requires 20
weeks, then a deployment requiring an Extract Layer to be devel-
oped might mean a 26 week project.
These percentages can be used to easily calculate a series of
indexes to compare
different types of
projects. For example,
if an index of 1.0 is
given to a project that
deploys one of the
pre-built analytic
applications against a
backend system where
100% of the compo-
nents above can be
used in their vanilla
state (such as Finan-
cial Analytics againsta typical implementa-
tion of Oracle E-
Business Suite), then a factor of 1.1 can be used for a project
where 50% of the Extracts must be modified (for example, a de-
ployment of Financial Analytics against a highly customized de-
ployment of Oracle E-Business Suite), and a factor of 1.2 can be
used when deploying one of the pre-built analytic applications
against a backend system where an extract layer has not yet been
written by Oracle engineers.
Whats the gain using Pre-Built
Applications versus BuildingCustom?
As described in the introduction to this whitepaper, the typi-
cal gain in projects which use the pre-built analytic applications
from Oracle, versus those that do not, is the difference of a factor
of 2.5 or 3.5 to 1. The diagram below demonstrates visually the
differences, and levels of effort typically applied to each compo-
nent of the application.
BI CONSULTING GROUP WHITEPAPER VERSION 2.0 FY08
- 2-
Copyright 2007, BI Consulting Group, LLC. All rights reserved. No part of this material may be reproduced, stored in a retrieval system or transmitted by anymeans, electronic, mechanical, photocopy, recording or otherwise, without written permission from the authors.
-
8/12/2019 BICG Buy vs Build
3/7
As-
sump-
tions
!This scope
and level of
effort is based
upon a de-
ployment of
the OBI EE
BI CONSULTING GROUP WHITEPAPER FY08
- 3-
PROJECT
METHODOLOGY
STAGE 1.0PROJECT INITIATIONSTAGE
STAGE 2.0DISCOVERY STAGE
STAGE 3.0DESIGN STAGE
STAGE 4.0BUILD STAGE
STAGE 5.0TESTING ANDVALIDATION STAGE
STAGE 6.0DEPLOY STAGE
STAGE 7.0MAINTENANCE ANDSUPPORT STAGE
Project MethodologyFor all projects, BI Consulting Group employs the following custom methodology which was written specifically for Oracle BI Enterprise
Edition projects. The methodology defines a seven stage process, with inputs (materials and information provided by the customer), tasks,
deliverables, and training recommendations identified at each stage.
Comparison of Project TypesIn the graphic below, we have summarized the relative level of effort typically required for
several different projects based upon the buy model (using Oracles pre-built analytic applications),
and have compared that to a build project (a project that requires the development of a data
warehouse built from scratch, against Oracle EBS, PeopleSoft, Siebel, JD Edwards or similar
systems). Details for each project type are outlined in the following pages, including a sample
timeline, duration of each stage in the project, typical team sizes, and typical assumptions.
Project Type
Buy
Build
1.2 Deployment of Pre-Built Applications against EBS, PeopleSoft, Siebel, or JD Edwards
1.3 Deployment of Pre-Built Applications against a non-Oracle backend
1.4 Custom Warehouse Build against EBS, PeopleSoft, Siebel, or JD Edwards
75 person weeks of effort
58 person weeks of effort
185 person weeks of effort
Copyright 2007, BI Consulting Group, LLC. All rights reserved. No part of this material may be reproduced, stored in a retrieval system or transmitted by anymeans, electronic, mechanical, photocopy, recording or otherwise, without written permission from the authors.
-
8/12/2019 BICG Buy vs Build
4/7
Assumptions
! This scope and level of effort is based upon a
deployment of the OBI EE Platform using one or more of
the pre-built applications, against an operational application
supported by the pre-built applications, such as Oracle EBS,
PeopleSoft, Siebel or JD Edwards.
! This scope and level of effort assumes that the
operational application has a typical number of
customizations and/or re-purposed data elements.
! This scope and level of effort is based upon 3 4
subject areas, and 40 50 reports, however, using BICGs
best practices design of reports and Dashboards, each
report module will have multiple views and selections,
resulting in dozens of permutations for each report.
Team Requirements
! Team Leader/OBI EE Architect
! BA/OBI EE Designer
! ETL Lead
BI CONSULTING GROUP WHITEPAPER FY08
- 4-
1.2 PROJECT SCHEDULE AND MILESTONES FOR DEPLOYMENT OF PRE-BUILT
APPLICATIONS AGAINS T EBS, PEOPLESOFT , SIEBEL OR JD EDWARDS
Copyright 2007, BI Consulting Group, LLC. All rights reserved. No part of this material may be reproduced, stored in a retrieval system or transmitted by anymeans, electronic, mechanical, photocopy, recording or otherwise, without written permission from the authors.
The duration below is based upon a 17 week project with three weeks of post production support. This timeline
sample is based upon the deployment of up to 3 - 4 subject areas across up to three Analytic applications (such
as Finance, HR, Supply Chain, etc.).
-
8/12/2019 BICG Buy vs Build
5/7
Assumptions
! This scope and level of effort is based upon a
deployment of the OBI EE Platform using one or more of
the pre-built applications, against an operational application
not supported by the pre-built extracts.
! This scope and level of effort assumes that the
operational application is in a relational data source, and that
the data model is available, and that data stewards exist to
understand the location of pertinent tables and fields.
! This scope and level of effort is based upon 3 4
subject areas, and 40 50 reports, however, using BICGs
best practices design of reports and Dashboards, each
report module will have multiple views and selections,
resulting in dozens of permutations for each report.
Team Requirements
! Team Leader/OBI EE Architect
! BA/OBI EE Designer
! ETL Lead
! ETL Developer
BI CONSULTING GROUP WHITEPAPER FY08
- 5-
1.3 PROJECT SCHEDULE AND MILESTONES FOR DEPLOYMENT OF PRE-BUILT
APPL ICATIONS AGAINST A NON-ORACLE BACKEND
The duration below is based upon a 19 week project with three weeks of post production support. This timeline
sample is based upon the deployment of up to 3 - 4 subject areas across up to three Analytic applications (such
as Finance, HR, Supply Chain, etc.).
Copyright 2007, BI Consulting Group, LLC. All rights reserved. No part of this material may be reproduced, stored in a retrieval system or transmitted by anymeans, electronic, mechanical, photocopy, recording or otherwise, without written permission from the authors.
-
8/12/2019 BICG Buy vs Build
6/7
Assumptions
! This scope and level of effort is based upon a
deployment of the OBI EE Platform and the development of
a custom datamart / data warehouse, against a back office
system, such as Oracle EBS, PeopleSoft, Siebel, JD Edwards
or similar systems.
! This scope and level of effort is based upon 3 4
subject areas, and 40 50 reports, however, using BICGs
best practices design of reports and Dashboards, eachreport module will have multiple views and selections,
resulting in dozens of permutations for each report.
Team Requirements
! Team Leader/OBI EE Architect
! BA / OBI EE Designer
! ETL Lead
! ETL Developers (2)
BI CONSULTING GROUP WHITEPAPER FY08
- 6-
1.4 PROJECT SCHEDULE AND MILESTONES FOR CUSTOM WAREHOUSE BUILD
AGAINST EBS, PEOPLESOFT , SIEBEL , JD EDWARDS OR SIMILAR SYST EMS
The timeline sample below is based upon a 34 week project with six weeks of post production support,
assessment, and planning for Phase II.
Copyright 2007, BI Consulting Group, LLC. All rights reserved. No part of this material may be reproduced, stored in a retrieval system or transmitted by ameans, electronic, mechanical, photocopy, recording or otherwise, without written permission from the authors.
-
8/12/2019 BICG Buy vs Build
7/7
About BI Consulting Group
BI Consulting Group (BICG) is Oracles largest consulting partner founded on Oracle BI
Enterprise Edition (OBI EE, formerly Siebel Business Analytics) and winner of Oracles
FY2007 Most Valuable BI Partner award. Our current track record includes more than
250 OBI EE projects, with more than 120 deployments of Oracles pre-built Analytic
Applications. Our staff comes direct from the Business Analytics division at Siebel and
Oracle, or from Enterprise Edition customers where they have lived and breathed Enterprise
Edition for years.
The average experience of our staff is 5+ years working on Enterprise Edition
implementations, and more than ten years in Business Intelligence and Data Warehousing.All that, and were still 100% referenceable.
BI Consulting Group provides our clients with experienced Enterprise and data
warehousing consultants who provide a variety of consulting services for the design,
development, deployment and maintenance of Oracle BI applications. These services
include project management; requirements development and application design; metadata
development; report and Dashboard design and development; data warehouse design and
development; ETL development; and Training through BICG University.
BI CONSULTING GROUP WHITEPAPER FY08
- 7-
HIGHLIGHTS
! Oracles largest consulting partner
dedicated exclusively to Oracle BI
EE with nearly 60 OBI EE experts
in the U.S.
! Collectively completed more than
250 Oracle BI Enterprise Edition
projects and more than 300 data
warehouses and datamarts.
! Average consultant has 5+ years
of Oracle BI Enterprise Edition
experience, and more than 10
years of BI/DW experience.
! Every customer is referenceable
including those who had previous
failures with other consulting
firms.
! Experts at both stand alone
implementations and deployments
of the Pre-Built AnalyticApplications from Oracle including
Financial Analytics, HR Analytics,
Supply Chain Analytics, and CRM
Analytics.
! BICG University teaches 10 key
OBI EE courses.
Please contact us at [email protected] we can be of any assistance in your
selection process.
Best regards,
Michael Martin
President, BI ConsultingGroup
Copyright 2007, BI Consulting Group, LLC. All rights reserved. No part of this material may be reproduced, stored in a retrieval system or transmitted by anmeans, electronic, mechanical, photocopy, recording or otherwise, without written permission from the authors.
Contact Us
Email: [email protected]
2130 Cliff Road
Eagan, MN 55122
www.biconsultinggroup.com
mailto:[email protected]://www.biconsultinggroup.com/mailto:[email protected]:[email protected]://www.biconsultinggroup.com/http://www.biconsultinggroup.com/http://www.biconsultinggroup.com/mailto:[email protected]:[email protected]:[email protected]:[email protected]