3. What Does A High Performance Team
Do?
Image source: http://hunterburgan.buzznet.com/photos/happybirthday/?id=53349941
What We Need
4. What You Say Your Team Goals Are
• Short Term (Cash Flow)
– Higher Income
– Lower Costs (Protect Revenue)
• Long Term (Year on Year Growth)
– Faster Time to Market
– Higher Quality Product
– Better Predictability/Transparency in Process
– Improved Team Morale
– Improved Customer Relationship
5. What You Think Your Team Goals
Are
• Short Term (Cash Flow)
– Higher Income
– Lower Costs (Protect Revenue)
• Long Term (Sustainability)
– Faster Time to Market
– Higher Quality Product
– Better Predictability/Transparency in Process
– Improved Team Morale
– Improved Customer Relationship
6. Your Actual Team Goals
• Short Term
– All Public Goals Met (On Paper)
– Nobody (Important) Gets Fired
– Nobody (Important) Quits
– Everyone (Important) Gets a Good Raise and Bonus
• Long Term
– Everyone Keeps or Increases Their Personal Power,
Starting with You, The Leader
11. Enterprise Change is
Usually Incremental
Incremental innovation:
Solution-driven creativity
Concrete practice
Extrinsic motivation
Needs an opportunity to contribute a solution
Radical innovation:
Problem-driven creativity
Abstract theory-related ideas
Intrinsic motivation
Needs time to work the problem
Radical and Incremental Creativity: Antecedents and Processes, by Gilson and Madjar, in Psychology of
Aesthetics, Creativity, and the Arts, 2011, Vol 5, No 1., cited in “Innovation Management Best Practices,” Jackie
Fenn , VP & Gartner Fellow, July 25, 2013
12. How do you direct the
driver AND motivate
the elephant?
19. Simplest Scaling: Opportunistically
Identified Reference Instances
Disperse the coached projects to get maximum
impact. (1 per major vertical, per location, for
example)
What you want for a reference instance:
Friendly, relaxed, and trustworthy sponsor
Small, non-critical, likely to be successful.
What you may get:
Desperate executive
Huge, mission-critical program already going up in
flames.
20. Further Out: Combination of New,
Repeat, and Self-Evangelizing
Business
Plan for staffing bump in coaches
(FTE is cheaper and stickier than FTC)
Teams generally like it once they
start. Keep counting them as they do
new projects
Teams can help each other without
you later.
23. Further Out: Measure
Actual Business ROI
Start driving from the business!
Project Planned/Actual Market Impact (Protect or
Expand)
Project Planned/Actual Operational Savings
Monetize “saved keystrokes”
Monetize labor cost savings with increased
internationalization
Planned/Actual Progress Towards Other Goals
Community support
Social Justice
26. Low Hanging Quality
Measures
Functional Quality measured by defect
trending
Performance and Security
Technical Debt Quantified: Total Cost of
Ownership
Unit Test Coverage
Code Complexity
31. Actual Low-Hanging
Practices
Collaborative, collocated requirements
workshops
Software environment virtualization
Refactored code base/attend to technical debt
Automated combinatorial-data-based testing,
supported by a robust test data strategy
Continuous integration, automated
deployment
32. Further Out: Seriously
Gritty Change
Collocation; team rooms
Staffing patterns the norm, not the exception, where all team
members are employed at the same time for the whole project
Planned, thoughtful and large-scale reduction of accumulated
technical debt
Empowerment of teams
Budget, vendor, contract, and hiring reform
Different methodologies to handle different problems within a
single company
34. What Are My Choices?
• Option 1: Keep Doing What You’re Doing (Lowest
Risk, Lowest Return)
• Option 2: Change Everything (Highest Risk, Highest
Potential Return)
-OR-
35. Acknowledge and Balance Your Team
Goals as a Portfolio
-0.2
0
0.2
0.4
0.6
0.8
1
-500000 0 500000 1000000 1500000 2000000 2500000 3000000
Net Present Value (NPV)
Size of bubble is proportional to total value generated over lifetime use of new
team techniques.
ProbabilityofSuccess
36. 1 Minute Theory of Portfolio
Tuning
Ensure long term return
Ensure short term team survival
Ensure steady flow of corporate recognition from “sure
bets,” balanced by potentially huge returns on high-
risk items.
37. Do you see a problem here?
-0.2
0
0.2
0.4
0.6
0.8
1
-500000 0 500000 1000000 1500000 2000000 2500000 3000000
Low Risk,
High Return
High Risk,
High Return
High Risk,
Low Return
Low Risk,
Low Return
Net Present Value (NPV)
ProbabilityofSuccess
38. Example Software Development
Transformation Goal Portfolio
Goal Cost Payback Time to Payback
Virtualize cross-
impact
environments in all
projects
Buy solution, learn
to use it (requires
cash outlay and
team time)
No need to buy
actual servers; no
down time for
testers
Same year: self
funding
Enforce code quality
standards (unit test
coverage, code
branching, size of
modules)
Slows development;
some short term
opportunity cost
Improves customer
morale; Reduces
total cost of
ownership 17% over
life of project
2-5 years
Launch all projects
with collocated
workshop
Travel costs 85% reduction in
scope creep; 15%
reduction in defects
6 months
Implement SAFe
framework
Completely disrupt
SDLC for all
All performance
measures improve:
business more
competitive
Minimum 3 years
39. Is there a difference
between stated and
unstated goals on teams at
sub-enterprise sized teams?
40. Agile By Numbers: KPI-Optimized
Software Transformation for the
Enterprise
August 6, 2013
Elena Yatzeck
http://pagilista.blogspot.com
Notes de l'éditeur
Simple, repeatable steps: governments, universities, fund-raising bodiesStructure, time, and messageEngage consultants, but still stay in chargeChange metrics versus BAU metrics
“Be Careful What You Wish For” is real. A team can only perform well with regard to a specific goal. So your first step is to determine what specific goals you want your team to solve together.
This is just the kind of slide you typically gloss over. But don’t. Your first step to putting together a high performing team is
This is just the kind of slide you typically gloss over. But don’t. Your first step to putting together a high performing team is
Agile for the regulated multinational is sometimes seen as something that is “just not right.” Discoveries: although corporations are not people, there are good people working for corporations. You are not being hired by a company. You are being hired by a person. Do you trust them? If not, don’t work for them.These big companies have big impacts, and it’s immoral to refuse to engage.Think of yourself as an anthropologist.
What has your experience been?Esther Derby’s agile excellence framework: values, mission & vision, structures & policyValues: Day to day actions are consistently aligned with core values?Mission and vision: priorities are aligned with values?Structures and policies: driven by value creation for customers, not task lists: This may be the area requiring greatest coaching.
SHOULD BE 4pm NOW. If not, slow down. If so, speed up.quick survey of materials supporting agile-at-scale to date.
Some of the “horror” people feel about agile at corporate scale is that it rolls out institutionally at first, not through intrinsic motivation and experimentation.
Aristotle: logos, pathos, and/or ethos: logic, emotion, or social pressure Stages for deploying ("quiet phase," delayed base-lining, executive-only meetings, team-only meetings, town halls and general announcements, scaling change)Dan and Chip Heath: Direct the Rider:- Find the bright spots- Script the critical moves- Point to the destinationII. Motivate the Elephant:- Find the feeling- Shrink the Change;- Grow your peopleIII. Shape the Path:- Tweak the environment- Build habits- Rally the herd
Exploring Transformational KPIs (50 min). Note that no matter what is on your dashboard, it is a proxy for the power structure of your organization.Exec level:Build the right relationshipsUnderstand the unwritten structures and motivationsControl word of mouth as well as published materialsDo not publish for a while (maybe 6 months)Take advantage of a “quiet phase” where you build up some success, and discover local successes.Then publish regularly, continuing to monitor the word of mouth
Team level: dashboard creates transparency of management practices, and perception of even playing field. Electronic data is more under your own individual control.
Another Dashboard!Executives still operate within an environment where speed, quality, transparency, and ROI get you promoted. So we can empower even middle managers this way.
Direct the Rider:- Find the bright spots: Baseline should call out all available successes in your definition of what is agile- Script the critical moves: Here is how we are going to measure. Keep your eye on the actual KPI- Point to the destination: DittoII. Motivate the Elephant:- Find the feeling: transparency for executives, authority for the team- Shrink the Change (goal is incremental change, not “total transformation”)- Grow your peopleIII. Shape the Path:- Tweak the environment—speaks to your definition of agile- Build habits--ditto- Rally the herd—celebrate successes
Start measuring lead and lag measures from the beginning using "facts," with a rule like “three statuses in any direction changes the to level sheet.”
Opportunistic is a key word here: in the enterprise, your efforts will initially be dwarfed by the challenge. This is not generally going to start as Initial projects should be considered “reference instances.” They should be chosen carefully to be influential and successful. You will want to choose small, non-critical projects. You may lose this battle, and instead find yourself doing a huge multi-continent project rescue. Just do what you can.
You need to decide the likelihood of success. The relaxed executive is likely to be less influential. You will need more “relaxed executive” successes than you will need big, dramatic program rescues.
SHOULD BE 4:15 NOW. If not, slow down. If so, speed up.This depends on the strategy of your organization. Needs to consider long term and short term ROI. Long and short term around employee retention and static quality of code. Know your strategy, and you’ll know what to graph here.
You will likely want to make initial promises around “throughput” not “reduced staff.”
This depends on the strategy of your organization. Needs to consider long term and short term ROI. Long and short term around employee retention and static quality of code. Know your strategy, and you’ll know what to graph here.
then graduate to measures which are more subject to mathematical transformations like "time to value" or "number of defects introduced into production." Note, this may work backwards from what you think: Use of surveys to gather lead information about customer and team satisfaction, then move to observations about actual team staff turnover and actual operational savings or revenue increases measured by business partners, etc.
Should be 4:30 now Here is the deal. If you’re doing a transformation, you are firmly at the “Shu” of “Shu Ha Ri.” You will most likely bring in an external coach to help you define the characteristics of the methodology for the group whose results you are going to measure. If you know where you want to go, in terms of what you want to see on your dashboard, then you can evaluate whether your consultant is supporting your business or not.
This is probably what you think. It’s definitely what your employees think.
These are all high cost in the culture of a large company:Off-shoring policies and “facilities” are tricky to tackle, plus physical changes unsupported by meaningful process changes with results can embarrass you.Switching the staffing pattern of 100 people or 1000 or 10000 is a serious proposition. Will require one-time hit to delivery. Easier for greenfield than for BAUTechnical debt changes require significantly more trust, because payoff is gradual and in the futureShow that one methodology works before you get people to tackle multiple onesCommand and control isn’t going to release its vise-like grip until the people in control see the benefits of letting the teams decideAdmin functions are their own deeply entrenched silos in big companies. They require a LOT of effort to change. You must have compelling local proof to do so.Middle management lives primarily to perpetuate itself, and it is the bulk of the power structure of a company.
Y axis indicates probability of success (0-1). X axis indicates net present value of each project. Size of dot indicates overall likely income from project over its lifetime.
Note this portfolio has no “low risk, high return items,” and one high risk, high return item. They’re avoiding high risk/low return except for one large-ish project.Other dimensions you might want to bubble chart: costs, time, return on investments towards different balanced score card values, so long as they are quantitative, regulatory compliance, operational cost, etc.
Corporations are not people. But people work at corporations. Find the ones you can work with, and help them. It’s all about leverage.
Simple, repeatable steps: governments, universities, fund-raising bodiesStructure, time, and messageEngage consultants, but still stay in chargeChange metrics versus BAU metrics