20121023 lkce 2012
TRANSCRIPT
Attaching Kanban to the Command & Control World of Project
Managers
Lean Kanban Central Europe 2012
[email protected]@nikolaus_rumm
Agenda
• Part I: The problem– What is project management ?– Why is it problematic with IT development projects ?
• Part II: Finding the right cure– What is organizational culture ?– How does it affect Kanban and project management ?– How to handle project management and Kanban in the same
organization ?
• Part III: Summary
The Project
A collaborative and temporary enterprise, frequently involving research or design, that is carefully planned to achieve a particular aim
Usually…• something singular• something new• something risky• something big
Source: Wikipedia, en.wikipedia.org/wiki/Project
Projects are temporary organizationsCEO
Procurement Delivery
IT Development
IT Operations
Customer Care
Sales
Domestic Sales
International Sales
Project Manager
Analysis Team Dev Team QA Team
• The project‘s organization is a down-scaled version of the company‘s organization• The organizational model is
the hierarchy• The idea behind it is damage
limitation (risk isolation) and decomposition• In fact it‘s a confession that
the company is not able to achieve some aim with the existing structures
Why they invented projects…
Hoover Dam Project
• Cost: $49 M (1936)• Planning time: 5 years• Implementation time: 5 years• Consortium size: 6 companies• Team size: 5,200
Manhattan Project
• Cost: $2 B (1946)• Planning time: 3 years• Implementation time: 4 years• Team size: 130,000
…and what we made out of this idea
Call Statistics Report Project
• Cost: $6,200
• Planning time: 3 weeks
• Implementation time: 2 days
• Team size: 0.5 FTE
CRM Maintenance 2013 Project
• Cost: $14,000• Duration: unlimited• Team size: 0.2 FTE
The awful truth is…• By 2012 the predominant IT
management model is the command & control bubble
• Everything is a project with a price tag
• What‘s not a project is called overhead and thus considered inefficient
• The majority of IT projects are just renamed cost units
• This is a sign of missing trust between upper and middle management
P1
P2
P3
P4
P5P6
P7P8
P9 P10
P11
The „weak matrix“ (release 2012)
Projects focus on Short Term Gain
• The idea of the temporary organization encourages optimization of short-term gain
• This includes accepting a customer rip-off, even if we lose the customer after the project has finished
• Some project managers focus on satisfying their line manager and not on the customer
The customer
The boss
Project manager
Do the least acceptable once he signed that contract
Impress her !
Project Managers tend to be Heroes
• Heroic management• Smooth operations are not good for
heroes• Possible trouble magnets• Prone to burnout in larger projects• Strong bias on dysfunctional
communication• Note that some employees actually like
that because the hero is responsible for everything
• „King of the hill“-syndrome– He who knows everything– He who plans everything– He who decides everything
Traditional PM requires Planning
The practise of project management often assumes that right at the beginning…
– The goals are clear– The goals are stable– The planner has complete knowledge
of the problem– The planner has complete knowledge
of how to achieve the goals
Each assumption is wrong
How wrong is your Plan ?
Requirements change at an average rate of 1-2% per month
Even if you know everything and elaborate a perfect plan a one-year project team develops 12-24% into the bin
Inability to follow the change in demand automatically implies reduced quality
Project Managers plan Input
• The planning/controlling object is the work package
• Work packages describe…– what to do– who will do it– when it will start– when it will be completed
• Work packages are input targets
Henry Laurence Gantt(1861 – 1919)
• Co-founder of scientific management
• Inventor of the Gantt chart (1910)
Input Planning
• Input planning focuses on what to do and not on what to achieve• Input planning is much more complicated and error prone than output
planning• Assumptions
– Each resource‘s productivity is known for each task– Each resource‘s productivity is stable
• Both assumptions are wrong• Productivity variation between developers is 1:10
Input vs. Output Planning
Project Goals
Project Goals
Product Vision
Product Vision
Features
Work Packages
Features
Product Increment
Product Increment
Ad hoc
Schedule Statistical Control
Output Target
Output Target
Output Target
Input Target
Input Target
Output
Output Target
Output Target
Output Target
OutputTransformation& Error Source
•Less work•Better quality
Why do they plan Inputs ?• Traditional project management is still
influenced by a mechanistic mindset from the industrial era (i.e. scientific management)
• Focus on efficiency and labour productivity
• Patterns– Find the best way how things are done– Eliminate variability by standardization– Divide labour until you can downskill the
worker
• These assumptions are not applicable to brain-workers
• All attempts to industrialize software development have failed
Frederick Taylor(1856 – 1915)
• Father of scientific management
Input Planning is not applicable to Software Development
Construction Project SW Development ProjectPlan Very detailed SketchyValidation of the plan Formal
IT support availableInformal
Dependencies between tasks
Very strict Medium
Variability of productivity Low HighFreedom of choice for the worker
Non-existing High
Changes Few ManyImpact of changes Local Possibly globalWork mode Follow the plan Follow the goal
(even if there‘s a plan)
The One-Shot Project Plan• Note that the principles of project
management require frequent replanning
• Many projects have no up-to-date project plan
• Reasons– Too much effort– Could surface the bad truth
• The plan is necessary for getting the approval, but after that no one is asking for it
• Outdated plans are a clear indication of cargo cult project management
• Common in organizations that treat projects as instruments of accounting and controlling and not for development
Organizational Culture
„How we do things around here to succeed“
• Quite stable• Very difficult (if you‘re the CEO) if not impossible (if you‘re an
external consultant) to change• Changing it takes at least 2-3 years• Attracts or chills possible employees
Source: William E. Schneider, The Reengineering Alternative
Assumption #1
All organisations are – fundamentally – living social organisms
• Stable core of beliefs, ethics and principles
• The system reacts to attempts that pull it off center with resistance and homoeostasis
• Fundamental changes are unlikely without a crisis
Source: William E. Schneider, The Reengineering Alternative
Assumption #2
Organisational culture is more powerful than anything else
• Agile implementations must consider strategy, leadership and – above all – culture in order to be sustainable• If the change is not
sustainable then the change initiative has failed
Culture
Leader-shipStrategy
Source: William E. Schneider, The Reengineering Alternative
Assumption #3
System-focused interventions work. Component-centered interventions usually do not.
– The system can’t be decomposed– Anything we do in agile
management must address the whole system
Source: William E. Schneider, The Reengineering Alternative
Assumption #4
Interventions clearly tied to business strategy work. Interventions not clearly tied to business strategy do not
– Start where you are– Accept how the company currently
works– The intervention must accomplish
the company’s purpose and not that of the change agent
– The others are always more
Source: William E. Schneider, The Reengineering Alternative
The Baseline
„If the management idea fits the nature of the organizational culture it will most likely work.
If not, it will most likely fail.“
• Kanban does not fit to all organizational cultures• Kanban is less invasive than Scrum, but still a big challenge
for most enterprises
Source: William E. Schneider, The Reengineering Alternative
Organizational Culture
The way we take decisions is…Personal
We pay attention to…
Impersonal
Presence and Reality
Possible Future
Goal Dominance
Method Gain control
Customer Relationship Controlled
Organizational Model Hierarchy
Archetype Military
Goal Synergy
Method Teamwork
Customer Relationship Cooperation
Organizational Model Team
Archetype Family
Goal Improvement
Method Personal growth
Customer Relationship Fulfilment
Organizational Model Network
Archetype Religion
Goal Excellency
Method Growth of expertise
Customer Relationship Exploit your USP
Organizational Model Matrix, ad hoc
Archetype University
„We succeed by being the
best“
Be the best !
„We succeed by getting and
keeping control“
Stick to the rules !„We succeed by working
together as a team“
Collaborate !
„We succeed by growing
people who fulfil our vision“
Grow !
Collaboration Culture Control Culture
Cultivation Culture Competence CultureSource: William E. Schneider, The Reengineering Alternative
The Six Kanban Practices
• Visualize work• Limit the work in progress• Manage flow• Make policies explicit• Implement feedback loops• Improve collaboratively, evolve experimentally
Source: Leopold, Kaltenecker, Kanban in der IT
Six Kanban Core Practices
The way we take decisions is…Personal
We pay attention to…
Impersonal
Presence and Reality
Possible Future
Collaboration Culture Control Culture
Cultivation Culture Competence Culture
Visualize work Limit the WIP
Manage flow
Make policies explicit
Feedback loops
Improve (based on models)
Scrum ?
Kanban focuses on operational excellence by getting statistical control and improving the system‘s culture
Core Processes of Traditional PM
The way we take decisions is…Personal
We pay attention to…
Impersonal
Presence and Reality
Possible Future
Collaboration Culture Control Culture
Cultivation Culture Competence Culture
Controlling
Closing
Initiating
Executing
Project planning
Traditional project management focuses on getting controlof a team and utilizing them to the highest possible extent
Finding the right Cure• Would you buy a medicine that
claims to help against all diseases ?• The six Kanban practices have to be
adopted to fit the organizational culture
• Kanban will fit control cultures, might be adoptable to collaboration and competence cultures and will most likely fail in cultivation cultures
• This is the same compatibility pattern as with traditional project management
„If the management idea fits the nature of the
organizational culture it will most likely work.
If not, it will most likely fail.“
Kanban Principles
1. Start where you are– Understand culture, strategy and leadership– Identify stakeholders, pain points and interests
2. Apply evolutionary, incremental change3. Initially respect job titles, roles,
responsibilities and processes4. Encourage leadership on all levels– Depending on the organizational culture this
might be difficult
Legacy ProductsLegacy
Products
Other Projects
Other Projects
The usual Stakeholders
ProjectTeam
Line Management
Customer
End Users
Other Projects
Legacy Products
PMO
Stick to the rules (but don‘t overperform)
Win the end users(quality, flexibility)
Win the customer(reliability, predictability)
Do not betray the line management (transparency)
Plan resources(availability)
Plan resources and build up creative tension
Happy customers will help you in keeping the management buy in
Win
Avoid
Join
Know your Strengths and Weaknesses(as compared to traditionally managed projects)
Strengths• Better control of productivity• Less variability• Usually better quality• Much more flexibility• Better integration with legacy
maintenance and other projects• Plenty of metrics available
Weaknesses• No stable plan by definition• Statistical control is not as sexy as direct
control (for traditional managers)• Some cultures might not like the idea of
a team taking decisions
Opportunities• Users/customers are generally
disappointed of traditionally managed IT projects
• Users/customers usually like the way of working with agile teams
• Long-term relationship with the customer is possible
Threats• You‘re a democratic alien in C&C• You‘re a heartless number cruncher in
collaboration cultures• You‘re mediocre in competence cultures• You‘re totally insane in cultivation
cultures• The PMO usually hates you• External interventions might erode your
agile implemenentation
Strategy 1: The eroding sandwich
• Perhaps the easiest way in established companies• Pros: easy, less conflict, potential growth in both directions• Cons: undermines end-to-end thinking, possible
concurrency with project manager, focuses too much on process mechanics
Traditional Project
Management
Customer IT Operations
Traditional Release Planning
Development (Kanban)
Strategy 2: Coexistence
• Perhaps the best way if supported by line management• Pros: exploitation of the full potential of Kanban possible, allows
the build-up of creative tension with traditional projects, prevents monoculture
• Cons: might lead to unproductive conflict, totally depending on line management‘s buy in, requires intense relationship management with line management, limited possibilities for cultural change
Traditionally managed project
Customer Line ManagementKanban project
Strategy 3: All-in
• Pure Kanban implementation; short term goal in startups, long term goal in established companies
• Pros: exploitation of the full potential of Kanban possible, advanced management/leadership principles possible
• Cons: maybe not the best fit for all customers (and their cultures) if you are an IT services company
Customer
Kanban project
Summary
• Traditional project management has an impressive track record but is not well-suited for IT development projects– Relies on a plan– Plans input, not output– Biased on waterfall– Projects tend to optimize short-term gain
• Traditional project management is about decomposition and getting control
• Many companies misuse projects for accounting
Summary (2)
• Organizational culture is at the core of a company and influences everything
• If the management idea fits the nature of the organizational culture it will most likely work; otherwise it will fail
• Organizational cultures can be divided in– Command – stick to the rules !– Collaboration – work together as a team !– Cultivation – grow !– Competence – be the best !
Summary (3)
• Kanban focuses on getting (statistical) control of a complex production system (short term goal) and improving it (long term goal)
• Project management focuses on getting control of a team and utilizing them to the highest possible extent (short term goal)
• Both methods have the idea of control, based on rules and rationale at the core
Summary (4)
• Kanban and project management fit well into command cultures
• Kanban and project management might work in collaboration and competence cultures
• Kanban and project management will likely fail in cultivation cultures (use Scrum there ?)
• Kanban and project management can coexist, as they ultimately are rooted in the same culture
The Kanban Community
Where do you think that the Kanban community (the ones that regularly meet at conferences like this one)
best fits into ?
The way we take decisions is…
Personal
We pay attention to…
Impersonal
Presence and Reality
Possible Future
Collaboration Culture
Control Culture
Cultivation Culture Competence Culture
Understand your bias and know your blind spots