Dancing With The Agile Goddess Agile 2008

04/07/23

Dancing With The Agile Goddess

Israel Gat

Agile 2008August 8, 2008Toronto, Canada

© Copyright 04/07/23 BMC Software, Inc.2

Agenda

What can Definitely be Accomplished - Agile Success in R&D

You Can Do It – A Secret Sauce for Level I Agile Enterprise

The Deeper Nature of Agile - Agile as a Dual Disruptive Methodology

Blueprint for Level II Agile Enterprise - End-to-End Agile

(Blueprint for Level III Agile Enterprise – Agile-Based-Market-Of-One

Benefits:– Customer-vendor interactions– Software creation and delivery paradigm– Nontraditional value chains)


Agenda





Blueprint for Level III Agile Enterprise – Agile-Based-Market-Of-One

Benefits:– Customer-vendor interactions– Software creation and delivery paradigm– Nontraditional value chains


› Two major BMC Software releases were assessed (BMC Performance Manager R2.3

and R2.4).

› Characteristics of the two releases:

› 90 – 95* full-time equivalent staff (FTEs)

› Started implementing the Scrum methodology in early 2005 (Waterfall

environment until then)

› Rally software tools and consulting

› Twelve Scrum teams coordinated through Scrum of Scrums

› 1,450 story cards

› ~1.4 million lines of new and modified code. (*) This figure applies to the BPM 2.3 and 2.4 projects; all in all BMC currently has close to 1,000

“Scrummers”

[This and the next six slides are based on the BMC/QSMA/Cutter Consortium study; August 2007]

The BMC/QSMA/Cutter Consortium Study


Industry Data from the QSM SLIM-Metrics Database

› Spans 20+ years

› Large heterogeneous database contains over 7,000+ projects

› Represents over 685+ million SLOC, 7+ million function points, over 600 languages, from 500+ organizations in 18 countries

› Adding 200 – 400 projects/year


Trendline Assessment – Schedule

Story Cards100 1,000

Mont

hs

1

100

10

Build Phase Time

Faster

12-13 months

4-5 monthsBPM 2.4

BPM 2.3

BMC is much faster than the overall industry trend:>95th percentile.BMC

QSM SLIM database

1 Sigma line style


Trendline Assessment – Schedule – Agile Sample

Story Cards10 1,000

Mont

hs

1

100

10

Build Phase Time

Faster

BPM 2.4

BPM 2.3

BMC much faster to market than the Agile samplefor >100 story cards releasesBMC

QSMA SLIM database

1 Sigma line style

Agile Companies

100


Trendline Assessment – Quality

Stories100 1,000

Bugs/

Defects

Bugs/Defects (Hardening Iterations)

10

10,000

1,000

100

Defects at normal levels in spite of fast speedBMC

QSM SLIM database

1 Sigma line style

BPM 2.4

BPM 2.3


Productivity Index Assessment

BMC

QSM SLIM database

1 Sigma line style

Agile Companies

Higher productivity in spite of running with more Scrum teams than other Agile projects

Story Cards10 1,000100

Productivity Index (PI)

0

35

15

25

20

5

10

30BPM 2.4

BPM 2.3

Productivity Index


Productivity Index (PI)Industry values by application type

0 2 4 6 8 10 12 14 16 18 20 22 24

Productivity Index (PI) w/ ±1 Standard Deviation

Business Information

Command and Control

Process Control

Scientific

System

Telecommunications

Engineering

Avionics

Microcode

Real Time

Real Time

2826


Agenda







The Secret Sauce of BMC’s Success

› The three ingredients of BMC’s secret sauce were:– The passion and pride of the craftsman– Creating a learning organization– Providing “air cover” for a long enough time for harnessing the Scrum methodology to our

specific environment/circumstances› Agile at BMC pretty much followed the evolutionary biologists “Punctuated Equilibrium”

paradigm advocated by Stephen Jay Gould - confluence of several factors that mature all-at-once:

– Long intervals of additive evolution (“Equilibrium")– “Punctuated" by short, revolutionary transitions, in which species became extinct and

replaced by wholly new forms– The Punctuated Equilibrium paradigm has successfully been applied to explain

phenomena in other domains, e.g. revolution in military affairs– It can/should happen in your company

› What BMC accomplished can be achieved by any fully committed company that has the leadership, know-how, flexibility and patience to let the Agile process evolve and mature

What A Company Needs for Agile Success


Sample Proof Points for Significant Agile Success

› BMC

› Exigen

› Follett Software

› IDX/GE Healthcare

› SirsiDynix

› Xebia


Agenda








Between Christensen and Corporate Inertia

Agile: a disruptive methodology on two dimensions

Interplay of 2 Dimensions


Agenda








End-to-End Enterprise Agility

Agile ComponentTeam

Agile Teams of Teams

Agile Enterprise

Scale (perpetual)

Measurement

A BC D

Requirementsrunway

IntentionalArchitecture

Enterprise Tooling

Iterate

Iterate (2 wks)

Design

Test

Build

Release(3-4 mo)

Hardenand Ship

Req

Plan

Iterate Iterate

Iterate

Agile Release Train

OrganizationalChange

OrganizationalChange

Agile Release Train

Requirementsrunway


3 Levels of Participation

I1 I3I2 I4 I6I5 I7 I8

Market Planning

Market/Solution Release PlanMarket/Solution Release Plan

Product RoadmapProduct Roadmap

Product Release PlanProduct Release Plan Product Release PlanProduct Release Plan

Direct Field ParticipationDirect Field Participation

Iteration Demos / Requirements BacklogIteration Demos / Requirements Backlog

3-4 monthsRelease Planning

I1 I3I2 I4 I6I5 I7 I8

2 weeksRelease Mgt


“xi” release

Marketing ReleaseTrain

“bsm” release

R1 60-90 days 60-90 days 60-90 days 60-90 daysR2 R3 R4 R5

Agile GTM Process – Roy Ritthaler’s Design

R&D Release Train

Engineering & Product Management Driven Frequency

Market Driven Frequency

“beta” release

“maintenance” release

Collaborative

Asynchronous

GTM Processes


Transition to Agile Organization

Agile R&DTeams

OperationalTeam

ExecutiveTeam

R&DTeams

Waterfall Optimized Organization Agile Organization

Executive Management


Agenda








Systemic Problems with Level II Agile Enterprise

› In a typical corporate value chain, change in velocity in any single link exposes inefficiencies in the subsequent link(s)

– PM Dev QA Marketing Sales Revenue Recognition Back Office Customer Support …

› Change in velocity of a few consecutive links often generates a catastrophe downstream

› Consequently, one encounters innate corporate drive to “harness” the disruptive Agile methodology

– Largely at a level lower than consciousness– Typically in the form of some kind of “executive governance”– “Collaborative governance happens in the trenches, among people who interact

everyday. The real governance solutions happen outside the executive suite…”[William Ulrich in “The Cutter Edge”; April 15, 2008]

› Bringing logic and reason to an interdisciplinary corporate balance which evolved over many years is not necessarily sufficient for success…

Sacred Laws of Software Engineering Revisited

› The BMC experience reported earlier in this presentation can be summarized as follows:

– Aggressive staffing: 2-3X the typical number of co-dependent Scrum project teams

– Racing against short deadline without incurring a hit on productivity– Gaining big time on productivity without incurring a hit on quality– (Distributed development environment)

› It seems that for this kind of Hyper-productivity in Scrum, two cornerstones laws of software engineering could be relaxed*:

– Brook’s law– Puttnam’s 4th law

› With some very far reaching ramifications with respect to the traditional release concept….

(*) I am deeply indebted to Michael Mah for this observation


Release-centric Agile Planning

Collect Customer Requirements

Establish a Prioritized Backlog

R1

R3

R4

R5

R6R7

R8

R9

R10

R2

R1

R3

R4

R5

R6

R7

R8

R9

R10

R2

Rele

ase

R5

R6

R9R2

Define a Release Based on Priorities

R1

R3

R4

R7

R8

R10

Back

log

time


Made-to-Order Releases

…Define a Release Based

on Priorities

Additional Customer Requirements

• “Specials” would be added in real-time

• “Specials” do not get “stuck” in the backlog

C2C1

Create Customer-tailored Releases

• Certain features, e.g. R9, get “demoted” to the backlog

R9Re

leas

e

R5

R6

R2

R1

R3R4

R7

R8

R10

Back

log

C-1

Rele

ase

R5

R6

R9R2

R1

R3

R4

R7

R8

R10

Back

log

R5R6

R2

R5R6

R2

C-2

Rele

ase

• The customer gets the “release” in 3-4 months max

Managing the Overall Project

• During an interim period; N special releases exist

• Once a year all special releases get consolidated to a standard release

R9

R1

R3

R4

R7

R8

R10

Back

log

Rele

ase

R5

R6

R2

C2

C1

Rele

ase


Agenda








Level Access/Ease Experience Price Product Service

III: Consumer seeks the Company

(Dominate)

Give me a solution; help me out in a bind

Establish intimacy with me by doing something no one else can

Be my agent; let me trust you to make my purchases

Inspire me with an assortment of great products I didn’t know about

Customize the product or service to fit my needs

II: Consumer prefers the Company

(Differentiate)

Make the interaction convenient for me

Care about my needs and me

Be fair and consistent in your pricing. I’m not necessarily after the lowest price

Be dependable in your selection and in stock position, so I can rely on you when I’m in a bind

Educate me when I encounter a product or a situation I don’t understand

I: Consumer accepts the Company

(Operate at Par)

Make it easy for me to find what I need, get in and out in a hurry

Respect me; treat me like a human being

Keep the prices honest; don’t jack them up or offer big savings when there are none

Be credible in your product and service offerings

Accommodate me; bend over backward sometimes to show me you care

Consumer Underworld

Block my way, hassle me, keep me waiting, make it hard for me to get in and out

Dehumanize me; disrespect me; ignore my needs

Be inconsistent, unclear, or misleading in your pricing

Offer me poor quality merchandise and services that I can’t use

Give me an experience I’d just as soon forget; me a reason to tell my friends and relatives to stay away

Agile-Based-Market-Of-One in the Context of Customer-Vendor Interactions & Relations

Source: Crawford & Mathews; The Myth of Excellence; New York, NY; Crown Business, 2001

What does Customer Intimacy Really Mean?


Agile-Based-Market-Of-One in the Context of the Software Creation and Delivery Paradigm

› Revolution in how we conceive, do and deliver the “sausage” to the top N customers– Disintermediation of the “filters” between the customer and the R&D lab– Delivery straight from the R&D lab to the user– Major value migration back to products/solutions:

• The economics driven by 150-400% attach rates through professional and ancillary services are turned around

Agile-Based-Market-Of-One Blows Traditional Value Chains to Bits


Continuing our Dialog

Israel [email protected]

www.theagileexecutive.com


mailto:[email protected]

http://www.theagileexecutive.com/

Dancing With The Agile Goddess Agile 2008

Technology

bmcqsmacutter

dual disruptive

vendor interactionssoftware

scrum methodology

secret sauce

deeper nature

delivery paradigmnontraditional

agile enterprise