LESS2010 Recap, Tuesday edition

I’ll be recapping the International Conference on Lean Enterprise Software and Systems. I plan to post updates during the course of the day.

Please note that the notes are primarily for my own use, though I hope they are of use to you too 🙂 . If I hear something that interests me I write about it.

Notes are regular script, my own commentary is in italic.

Ps. You can follow the conference on Twitter by searching #less2010

Keynote: Professor Duc Truong Pham

Beyond Lean (Fit Manufacturing) paradigm

Based on Agile, Lean & Sustainable manufacturing

Lean (manufacturing)

• Lean is good for high volume low variety situations
• There’s a risk of becoming too Lean

Agile (manufacturing)

• Agile is about reaction time, flexibility, speed
• Agile enables getting the product to market fast and to respond to change
• Agile is good for low volume high variebility situations

Agile manufacturing concepts

• rapid prototyping
• virtual prototyping
• E-manufacturing
• CAD/CAE
• etc.

Concurrent engineering

• Parallel processes of design / development phases
• Time is saved

Potential problems in Agile

• over-reactive – unstable
• over-flexible – inefficient
• over-engineered – complex
• over-expensive – unsustainable

Sustainability

• not only environmental issues
• not only energy issues
• not only social issues
• economically sustainable manufacturing
• …something that can continue a long time
  • Long time: Stora Enso since 13th century, Stella Artois since 14th century, Elsevier since 16th century

Built to Last

• Clock builders, not time keepers
• Core preservation (values and ideology)
• Cult-like cultures
• Opportunistic experimentation
• Home-grown management
• Continuous improvement and innovation
• etc.

In Search of Excellence

• Get on with it (compare to Customer Development: go outside)
• Learn from the customer ( -“- )
• Foster innovation
• Nurture champions
• Treat employees as the main source of productivity gains (lean sw dev)
• Hands-on, value-driven mgmt philosophy
• Few admin layers, few people on top
• Core values but individual autonomy

Fitness

A Fitness Manufacturing company (FMC) has the qualities of agile manufacturing, lean manufacturing and sustainable manufacturing (to put it short)

FMC must

• know its balance and imbalance points
• make use of leverage
• control growth
• strive for steady state
• minimise entropy
• maximise simplicity
• etc.

Lean is mean,

Agile is Fragile,

Fit is sustainable.

Hmm.

Pham’s message for software industry: if you are doing Lean or Agile, add sustainability to it. Make sustainability explicit. You may already do things that make the company sustainable but you need to make it visible and deliberate.

Honkonen & Eskelinen: Asking Questions, Coaching

Excellent stuff. I’m going to devote a separate blog post on this one.

Keynote: Juhani Risku

Lean & Agile is paradise

Creative manager 1 in 6,7 billion

How to manage a 50000 person company as a small workshop?

• There you need agile & lean

Beautiful things:

• Google Earth
• Segway

Strange things:

• Brands, not realistic at all, a promise of something
• There are failures and mistakes in the promise
• Is brand marketing or the real thing?
• Only way to transform is to rearrange all the structures – that’s why brands need to be under creative

Leadership Structure

• Why so many financial executives in the board of a creative company?
• Are they financial institutions?
• Ideal: Chief Drama Director, Chief Future Director, Chief Community Officer, Chief Creative officer
• Only thing needed from the financial officer is the answer to the question “Do we have money?” then he can be dismissed and called back after some time for the same question

Slogans

• Less is more
• Don’t reinvent the wheel
• This is not rocket science

All of these are wrong – everything is the opposite

Original meaning of less is more – less outdated ornaments so that we can have more modern stuff (Ludvig Mies van der Rohe)

Ask the user

• No car company has asked the user what they needed prior to designing
• it should be don’t ask the user
• users wouldn’t have come up with Segway
• users are useful when you are developing, not when you are creating
• Never ask the user in design-based business

Innovation?

• CIICPDIL – the whole process
  • conception – idea- innovation – concepting – prototyping – design – implementation – launch

Concepts are the most important assets

• more important than people or money
• Concepting is more important than design
• Industrial designer for iPhone O – it’s there! (the form)
• You have many ideas – concept has the dramaturgy, but enough openness
• Managers usually do not recognise the winning concepts

Gates – Approval

• People who don’t have knowledge get into approval positions
• They don’t have the courage to leave
• They just stall

Scaling concepts

• Filtering 50000 – 5000 – 500 – 50 – 0 (down the process)
• Management didn’t understand about systems

Customer experience

• Most dangerous thing in a design-based company
• Everybody can have training and opinions on what is the customer experience like

Having many strong leaders is better

• 1000 people – 1 leader
• 50000 people – 10 leaders

A Camel is a Horse Designed by Committee

Post-it innovation – c’est passée et stupid…

• not the user but personnel as designers

Analyze classic products such as cameras

• go back with the design ideas to the history

Most important subjects:

• Morphology is understandability
• Criticism is caring

New mathematics, how come?

• What is the formula of beauty?
• How to go to USA with a product?
• New mathematics for me could be systems thinking, simulation and modeling

Creative organization

• Nokia had a community of 3,5 million app developers
• They were not told what to create
• So they created 50000 calendars
• The phone already had one

Creative organization (how to)

• Flat organization NO!
• Direct organization NO!
• Motivation NO!
• real answer is complex and takes years to collect
• no one is teaching it

Future and patterns

• Search relevance
• Background processing
• Profiling
• Understandability
• Structuring knowledge
• Information visualization
• Genius Loci

Google is arrogant

• Does first, goes to court later
• Pushes things
• Does not care about what everyone else things
• One king decides
• This is good
• Other kinds of companies do not do that well

The winner is

• the Quickest and the Boldest
• University or Company
• not Aalto University

They need to Manage, Lead and Orchestrate

Markus Andrezak and Bernd Schiffer – Kanban & Technical Excellence

Another thing I need to get back from. Interesting story conveyed as a play, actually.

Christian Blunden – Collaborative Lean Delivery

I’m sorry to note that this recap is basically a copypaste operation of the text on the  speaker’s slides

Clean delivery (catchphrase)

Context:

• Some info about kanban
• Client in a hybrid agile model
• departmental silos
• projects measured by cost
• late cycle quality – iterations with quality added at the end
• late heroism
• break and fix mentality
• gated sign-off
• project bonuses for upper management
• 3rd party middleware dependencies

Problems faced

• out of date specs
• rework – walls full of red defect cards
• tossing work over the fence
• balancing throughput between silos
• communication and handovers
• unnecessary agile ceremonies
• off-shoring model
• delays from external suppliers
• absent stakeholders

Solutions found

• Clean, collaborative lean
• Radical change
• Continuous improvement process
• Micro inception (inception = flash requirements gathering process) during sprint 0 (spec as a source, but start from scratch)
• a new story wall (in dev, in qa, in spec => in process), get rid of silos
  • build quality in, no separate QA
  • pull for analysis where needed
  • no WIP limit as the stakeholder was the bottleneck
  • expedite category in the in process section
• single story flow
  • one miniteam per story
  • one story per miniteam
  • some extra time for some team members
  • microiteration process with acceptance criteria
• work slower to get quicker
  • no defects for things in process
  • defects for items in process are fixed straight away
• threesomes
  • 2 devs and 1 QA
  • removing the silo between devs and QA
  • same thing with the business analyst
  • so a foursome?
  • learning improved for the team
  • quality of acceptance criteria was better
  • far fewer bugs
  • quicker iterations
• Slacktime
  • time for non-feature related improvements
• Continuous showcases
  • no sprint review at the end of sprint
  • just-in-time showcases when stories were completed
• story huddles
  • delayed the requirements spec to the point when the story was picked up
  • information was as up-to-date as possible
  • led to less rework
• continuous retrospective
  • retrospective wall
    • “idea box”
  • were discussed in the next standup meeting
• suppliers are partners
  • relationship was not strong at the start
  • direct connection with the devs in partner companies
  • they were given the integration test set to run on their own

Further experience

• different team = different constraints

Conclusions

• success lived on
• obvious measurements: a lot of projects failing before, project success after
• prepare for debate
• don’t de-humanize
  • retrospectives are important, you cannot reduce them to feedback boxes
• this was an example of pushing a practice or paradigm
  • is it ethical?
  • asked for forgiveness, not permission

Jurgen Appelo

Slides to Jurgen Appelo’s presentation are available from his blog.

Single Piece Flow in Kanban: the Software Work Cell – Arlo Belshee

A risk in kanban software projects – a rising number of boxes for different kinds of in process states for cards.

One reason for this: from scummerfall to kanbannerfall.

The right way to do it:Toyota Production System + XP (technical practices) + kaizen

Visual of the board:

• two items WIP (work in progress)
• additional notes containing more information about the two notes (tests, analysis etc.), moving when work has been done
• now (WIP) and later sections (vision, queue)
• detective board comparison
• vision slot is for product vision

Process control

• the wait time for item to get done from the time it is posted to the queue
• Is enough for other stuff than special ones (tradeshows etc.)

XP practices

• co-located team
• 100% * pair programming
• rapid feedback
• “change your organization or change your organization”
• These are about communication and learning
• Which are important in single piece flow

Continuous, Incremental Design & Architecture

• Agile transformation typically works for 1 – 1,5 years
• after that there’s enough technical debt to slow development
• if system that is designed at sprint 0 is extended ex-temporenously, you can end up with not the ideal system
• that is if XP practices of just-what-you-need and refactoring are not used
• there probably isn’t such a thing as too much refactoring
• good design: everything is named well and is testable

Continuous, Incremental Testing

• 4 kinds of errors: porgramming, reqs, design, process
• TDD, ATDD is not enough

Continuous, Incremental Deployment

• lots of ways to do it
• no single solution

Starts with generalizing specialists