Discovery Kanban @ LKSE15

Discovery Kanban

Patrick Steyaert

Knowledge work

3

Exploiting existing

knowledge

Exploring new

knowledge

Finding out which problems to solve and creating solutions for problems

that have not been solved before

Using past experience and knowledge to create solutions for known problems and situations

Solving hard problems, including the problem of finding out what problems to solve (or opportunities to capture)

Delivery Discovery

Balancing Discovery and Delivery - © Patrick Steyaert, 2015

Software development is knowledge work

Ref.: Allistair Cockburn

4

Users &Sponsors

Testers BusinessAnalysts

Programmers

UI Designers

Detailed decisionsabout function

and data

Decisions about program

structure

Detaileddecisions about

externalappearance

Decisions about function

and style

Decisions about systemcorrectness

The knowledge discovery process

§  Features

§  Change requests

§  Fixes to production defects

§ …

Delivery

5


Discovery

Agile Business Requirements training - © Patrick Steyaert, 2015

Viability

Desirability

Feasibility

6

§  In the ideal world, work flows with little friction

§  In the “messy” real world there are all kinds of

friction: silos, specialists, communication, resistance to change, …

§ Kanban is a way of dealing with the present (friction in your current reality) and connecting with the future (flow in the ideal)

Kanban helps to deal with the real world

7


Continuous change and innovation 8

Future

Present

Dealing with the present & connecting with the future

Digital Customer value

Networked Pull

Analog Shareholder value Hierarchical Push


Shared visual understanding 9


Kanban principles and practices 10

Ideas Reday for Dev

Development Testing Ready for UAT Done

ongoing done 5 ∞

Abandoned

5

3

A user story

A bug

A user story

A user story

A user

story A user story

A user story

A user story

A user story

A bug A user story

A user story

A user story

A user story

•  Visualize

•  Limit WIP

•  Manage Flow

•  Make Process Policies Explicit

•  Implement Feedback Loops

•  Improve Collaboratively, Evolve Experimentally (using models/scientific method)

•  Start with what you do now

•  Agree to pursue incremental, evolutionary change

•  Initially, respect current roles, responsibilities & job titles

Core practices Foundational principles

Ideas Ready for Dev

Development Testing Ready for UAT Done

ongoing done 5 ∞

Abandoned

5

3

A user story

A user story

A user story

A user story

A user story

A user story

A user

story A user story

A user story

A user story

A user story

A user story

A user story

A user story A user

story A user story

A user

story

Delivery kanban - focus on “work”? 11

A user story

WORK items

Limit WORK in progress

12


Knowledge work (at scale)

information

decisions options

More than work alone!

How does this affect flow?

Delivery

Discovery

Exa

mple

Dis

cove

ry K

anban

13


Idea (>15)

Concept (>10)

Specification (>6)

Fragmented ideas

Upstream (option) Kanban - discovery precedes delivery -

Coherent specifications

Options, selection and commitment 14


Options Selection Commitment

Convergent process

Mor

e th

an j

ust

visu

aliz

atio

n

15


Idea (>15)

Concept (>10)

Specification (>6)

Minimum limits to ensure sufficient

options to downstream process

Upstream (option) Kanban - discovery precedes delivery -

FLOW

Managerial approach

Given goal

Ava

ilab

le m

eans

M1

M2

M3

M4

M5

Selecting between given means to achieve a pre-determined goal


16

WHAT HOW WHY

17


“Why”, “What” and “How”

can be known upfront

Discovery precedes delivery

Post-Hypothesis (known known)

Not all knowledge work is alike 18


Post-Hypothesis Known known

“Why”, “What”, “How” can be known

upfront

Discovery precedes delivery

Hypothesis Known unknown

“Why” is defined “how” is discovered

Discovery and delivery alternate

Pre-hypothesis Unknown unknown

“What” is known; “why” and “how” are

discovered

Delivery precedes discovery

Hidden assumptions Unknown known

Observation as a basis


19

“why” is defined

“how” is discovered

Discovery and delivery alternate

Hypothesis (known unknown)

Idealistic approach

Vision Now Next target

“How” is discovered

“Why” is defined

In the idealistic approach, the leaders of a project set out an ideal future state that they wish to achieve, identify the gap between

the ideal and perception of the present, and seek to close it. Dave Snowden

20


Applying the scientific method

Design and run experiment

(DO)

Create hypothesis

(PLAN)

Study results (CHECK)

Evolve model (ADJUST)

Agile Business Requirements training - © Patrick Steyaert, 2015

21

Exam

ple D

iscovery Kanb

an

22


Backlog Plan Do Adjust 3 2 3

Experiment Kanban (1) - Discovery and delivery alternate -

Check 2

FLOW

Hypotheses

Limit experiments in progress

Explicit feedback loop (internal or

external)

Validated learning 23


Ideas

Build

Product

Measure

Data

Learn

Hypothesis 24

We believe that <Building this feature> <for these people>

Will achieve <this outcome> We will know we are successful when we see

<this signal or measurable result>


Exam

ple D

iscovery Kanb

an

25


Backlog Build Measure Learn 4 3 2 3

Hypothesis to be validated

Experiment Kanban (2) - Discovery and delivery alternate -

Options that can be exercised

Limit the number of un-validated assumptions

Creating options 26


Exercised Option

Hypothesis Option creation exercising

Down payment House price will go up House

MVP (option to pivot or persevere)

User need Full product development

Divergent process


27

“what” is known

“why” and “how” are discovered

Delivery precedes discovery

Pre-Hypothesis (unknown unknown)

Entrepreneurial approach 28


Given means

Imagine possible new ends given available means

E1

E2

E3

E4

Imagined ends

M1 M2

M3

M4

M5

WHAT HOW WHY

§ Focus on what you can do and do it, without worrying much about what you ought to do

§  Begin with who you are, what you know, whom you know

§  Immediately start taking action and interaction with other people

§ Goals and network concurrently converge §  People you interact with self-select into the

process

§  Each commitment results in new means and goals

§  As resources accumulate in the growing network, constraints begin to accrete that reduce possible changes in future goals and restrict who may further join the network

Effectuation

29


Exam

ple D

iscovery Kanb

an

30


Backlog Plan Do Adjust 4 3 2 3

Actions

Action Kanban - Delivery precedes discovery -

Check 2

Limit the number of actions in

progress

Not really a feedback loop, just verification

It looks like PDCA, but it is not really PDCA!


31

Observations as a basis for discovery

Hidden assumptions (unknown known)

Path dependence The set of decisions one faces for any given circumstance is limited by the decisions one has made in the past, even though past circumstances may no longer be relevant.

32


Options

Emerging path

Lock in

Vari

ety

of o

pti

ons

(man

ager

ial flex

ibili

ty)

Critical juncture

Definition from Praeger, Dave. Our Love Of Sewers: A Lesson in Path Dependence

Pre-hypothesis Hypothesis Post-hypothesis

Weak signals – ugly babies & invisible gorillas 33

Every movie the company makes starts out "ugly”; Ill-defined ideas need protection the most, lest they die too young. - Pixar president Ed Catmull

Our minds don't work the way we think they do. We think we see ourselves and

the world as they really are, but we're actually missing a whole lot.

- Christopher Chabris and Daniel Simons


Observe-Orient-Decide-Act

Boyd’s strategic decision making loop

Observe

Orient

Decide

Act

Exam

ple D

iscovery Kanb

an

35


Backlog Observe Orient Act 4 3 2 3

Orientation Kanban - Observation as a basis for discovery -

Decide 2

Strategic decision cycle: Observe-Orient-Decide-Act

FLOW

Limit the number of decisions in

progress

Observations (decisions to be

made)

All together now – the discovery cycle 36


Upstream (option) kanban

Experiment kanban

Experiment

Vision

Known unknown

Constraints

Trigger

Unknown unknown

Orientation kanban

Action kanban

Model

Options

Known known

Commitment

Weak signal

Unknown known

Orientation

All together now – Program Management Kanban

Visualize “actions” and “experiments”

Visualize “decisions to

be made”


37

>4 5 4

>4 5 4

Visualize “commitments”

Visualize “options”

All together now – Product Management Kanban


38

Visualizing knowledge work 39


*oobeya = project room

Options Commitments

Observations Actions

Experiments

Not all Kanban is alike

Delivery Kanban Discovery Kanban Visualize work

Limit work in progress

Manage flow of work

Organize feedback loops

Evolve experimentally

Visualize decisions and options

Limit unvalidated assumptions and decisions in progress

Manage flow of decisions and options

Explicit feedback loops, customer feedback loops

Mixed change

Not all Kanban is alike

Delivery kanban Discovery kanban

Improving fitness for purpose of service delivery

Improving the fitness for purpose of organizations in a continuously changing landscape

Demand is established and there is more demand than capability

Not just delivery against established demand but also dealing with uncertain demand

@PatrickSteyaert

[email protected]

References 43


Discovery Kanban @ LKSE15

Business

user story delivery

user story work items

delivery patrick steyaert

production defects delivery

work flows

kanban principles

ideal kanban

known problems