Kanban for���Portfolio Management

Gaetano Mazzanti @mgaewsj

Agile42

sounds familiar?

“too many things to do” “slipping schedule”

“continuous resource reallocation, multitasking”

“too slow to react”

demand

capacity

the illusion of predictability

delays

defects plan

uncertainty emerging knowledge

impediments

more detailed planning rigid change management

additional overhead

welcome to uncertainty

Project Portfolio Management (PPM) is the centralized management of processes, methods, and technologies used by project managers and Project Management Offices (PMOs) to analyze and collectively manage a group of current or proposed projects based on numerous key characteristics. The objectives of PPM are to determine the optimal resource mix for delivery and to schedule activities to best achieve an organization’s operational and financial goals ― while honouring constraints imposed by customers, strategic objectives, or external real-world factors.

wikipedia

traditional portfolio tenets

Governance

Compliance

Control

PMO (budget administration, staffing, auditing)

traditional portfolio “pillars”

Top-down

Centralized Annual Budget

Detailed Planning

Fixed date, fixed cost, fixed scope���(fixed fate L)

Cumbersome electronic tools

starting vs finishing

people as “resources”

full capacity allocation

PUSH!

we love Excel!

rebooting teams

project “completed”: ���reboot the team and…���

start again L ouch!

teams are rebooted, project to project, putting together different people each time and expecting to have predictable outcomes

Tuckman model of team development

On time, On Budget & Conformant to Specs

oops!

how is success measured

inspired by Joakim Sunden

What is Lean Kanban?

The Kanban method is an���evolutionary approach to change

Lean

add value without generating waste

muri, mura, muda

principles of the���Kanban method

start with what you do now

agree to pursue incremental, evolutionary change

respect the current process, roles, responsibilities & titles

encourage acts of leadership at all levels in your organization – from individual contributor to senior management

Kanban���six core practices

visualize ���limit WIP���manage flow ���make process policies explicit���implement feedback loops���improve collaboratively, evolve experimentally

if you can’t see it���you can’t manage it

Visualize

Knowledge work is invisible. If you can’t see something, it is (almost) impossible to manage it.

Collaborative process improvement is almost impossible if work and workflow are invisible and if process policies are not explicit.

David J Anderson

a Kanban board

visualize to understand the process

how the process works and fails to work for you and your customers

how it can be changed

expose dysfunctions

Limit WIP

balancing���demand & capacity

demand

capacity

no matter how good you are at filtering���capacity is always full (and more)

# of items in system (WIP)

B

C D F

E

A

ideas elaborate do delivered validate

G

H

I

# of items in a specific state = 2

# of items in���specific area = 4

total # of���items = 9

limiting WIP

fewer commitments

more options

finishing stuff-> stop starting start finishing

100% utilization myth idle work

vs

idle workers

efficiency (doing things right)

vs

effectiveness (doing the right things)

multitasking sucks

it’s easier to start than it is to finish

multitasking sucks

visualphotos.com

123456

YOUR MULTITASKING ���IS MY BOTTLENECK

Jim Benson

12

34

56

let work flow let work flow

visualize flow

ideas elaborate do delivered validate

B C

D

F E

A

visualize flow

B

C

D

F E

A

ideas elaborate do delivered validate

highlight issues

B

C

D

F E

A

ideas elaborate do delivered validate

!

!

measure flow

B

C D F

E

A

ideas elaborate do delivered validate

G

H

I

cycle time

= WIP / throughput

(throughput = average completion rate)

Little’s Law

flow = speed * density

slack (%) absorb variations

% capacity utilization

queue size

queue size grows

exponentially at high capacity 0  

5  

10  

15  

20  

25  

0   10   20   30   40   50   60   70   80   90   100  

make process policies explicit (i.e. definition of done, WIP limits, pairing reviews, test automation, …) implement feedback loops (i.e. standups, retrospectives, operation reviews, metrics, …) improve collaboratively, evolve experimentally (using models and the scientific method)

Lean Agile Portfolio traits

Value driven

WIP limited

Decentralized, adaptive

Empirical, Probabilistic (statistical data support)

Lightweight business case/specs

Lightweight estimation

Physical Boards

Traditional Portolio Management

Centralized Annual Budget

Detailed Planning

Fixed date, fixed cost, fixed scope���(fixed fate L)

Compliance driven, top-down approach

Cumbersome electronic tools

Kanban board design Ideas

ooBiz Case Tech

Assessment15-30 5-12

Committed5

In Progress5

Verification5

RejectedCommitment

Point

Busines

s

R&D

source David J Anderson

Kanban board designs

source Pawel Brodzinski

demand analysis

where does work come from?

how is value assigned (and verified) to initiatives?

is arrival rate predictable? (usually not)

Ideasoo

Biz Case TechAssessment

15-30 5-12Committed

5In Progress

5Verification

5

RejectedCommitment

Point

Busines

s

R&D

work item types

standard

fixed date

urgent/emergency

intangible

…

small

medium

large

…

growth

maintenance

innovation

…

the magic of visualizing

waiting for approval initial design,

planning, outsourcing,

mold design, etc.

demand

capacity

one week later

a new board

color = team

46  

flow

(bad) flow

Z Z ZZ Z

Z Z ZZ Z

Z Z ZZ Z

bottlenecks

B

C D F

E

A

ideas elaborate do delivered validate

G

H

I

gate

J

K

bottlenecks

B

C

D

F

E

A

ideas elaborate do delivered validate

G H

I

gate

J

K L

bottlenecks

B

C

D

F

E

A

ideas elaborate do delivered validate

G H

I J

K

L

gate

items queueing up

empty area

limit WIP

by # of initiatives (size matters)

by size

by team capacity

by space available

by conversation

…

limit WIP by���team capacity / space available

mar apr may jun jul aug sep

TeamA

TeamB

TeamC

TeamD

TeamE

once again: ���complexity and uncertainty are

intrinsic to knowledge work

we must be able to���react to unfolding events and���adapt to changing circumstances

David J Anderson

but… we hate uncertainty

we prefer to take a decision now

even if it risks being wrong

i.e. annual budget (we remove options, we increase risks)

decision making in uncertainty

balance:

risk/return

maintenance/growth

short term/long term

etc.

selection

what to do next?

what to finish next?

last responsible moment selection

Discoverwhat to do

Do It Validate Outcome

Reject Commitment

keep your options open

pull

selecting/pulling���a project

do first

do next

do soon /���re-evaluate

do sometime or reject

fit with goals

timeliness (urgency, ���

Cost of Delay, ���fit with strategy)

source GDS select what you don’t do

what about���specs & estimations?

what if you have 50 candidate initiatives and only 15 can be done?

what’s the point (cost) of specifying and estimating all 50?

precision is worthless

guesstimation -> range + historical data

check and adapt

when starting a project

we simply need a rough idea of size and���an understanding of (un)certainty

which parameters should you consider to select what to do next

ROI, IRR, NPV, etc. ?

Do you know your Cost of Delay???

ROI Cost of Delay���

(1 month)

Project A 25% 5k

Project B 15% 30k

which one would you select next? example credit: Kenny Rubin

cost of delay &���classes of service

cost

time

cost

time

cost

time

classes of service and WIP standard work = 60%

urgent = 10%

due date = 20%

maintenance = 10%

classes of service, WIP, expedite lane

D

H

J I

F B

A E C

6

M L K 2

1

1

O

Q

N

P

EXPEDITE LANE

G

ideas elaborate do delivered validate

metrics?

try

measure learn

goal

measure learn

as a learning loop

control chart

4

6

8

10

12

14

16

average cycletime

UCL

LCL

cycle time

knowledge work has high variability

time!

CFD���Cumulative Flow Diagram

monitoring progress 90% complete

remaining 10% could take forever

is usually meaningless

90% of what?

burn-up chart

time

outcome

enough…

Lean Agile Portfolio traits

Value driven

WIP limited

Decentralized, adaptive

Empirical, Probabilistic (statistical data support)

Lightweight business case/specs

Lightweight estimation

Physical Boards

Traditional Portolio Management

Centralized Annual Budget

Detailed Planning

Fixed date, fixed cost, fixed scope���(fixed fate L)

Compliance driven, top-down approach

Cumbersome electronic tools

Portfolio Kanban���is more about���

changing/transforming the organization���than about���

stickies on a board

Mike Burrows

no change disruptive change

Kanban

inspired by the works of many, including���Mike Burrows, Pawel Brodzinski and David J Anderson

Kanban

learn from the people plan with the people begin with what they have build on what they know

Lao-Tzu

Gaetano Mazzanti

@mgaewsj

gaetano.mazzanti@agile42.com

Agile42

Kanban