Metrics In An Agile World

Editor's Notes

• #2: 1000 colored 3x5 cards or sticky notes, tape/tacksSnacks?, pens, notepads.STAGE: Flip charts? Whiteboards?
• #3: In Shanghai, course on Agile for TestersOne group didn’t seem to be fully engaged – not happy, not comfortable.Root Cause Analysis exercise
• #4: Cultural/linguistic? Had I not explained RCA well?No, they were serious. This was their greatest pain-point.They had actually tried 5 Whys, but none of us was happy with THAT!Through conversation, we rewrote it <click>
• #5: operant conditioning: A man went fishing one day. He looked over the side of his boat and saw a snake with a frog in its mouth. Feeling sorry for the frog, he reached down, gently took the frog from the snake, and set the frog free. But then he felt sorry for the snake. He looked around the boat, but he had no food. All he had was a bottle of bourbon. So he opened the bottle and gave the snake a few shots. The snake went off happy, the frog was happy, and the man was happy to have performed such good deeds. He thought everything was great until about ten minutes passed and he heard something knock against the side of the boat. With stunned disbelief, the fisherman looked down and saw the snake was back with two frogs!"What Gets Rewarded Gets Done," by Michael LeBoeuf
• #6: operant conditioning: A man went fishing one day. He looked over the side of his boat and saw a snake with a frog in its mouth. Feeling sorry for the frog, he reached down, gently took the frog from the snake, and set the frog free. But then he felt sorry for the snake. He looked around the boat, but he had no food. All he had was a bottle of bourbon. So he opened the bottle and gave the snake a few shots. The snake went off happy, the frog was happy, and the man was happy to have performed such good deeds. He thought everything was great until about ten minutes passed and he heard something knock against the side of the boat. With stunned disbelief, the fisherman looked down and saw the snake was back with two frogs!"What Gets Rewarded Gets Done," by Michael LeBoeuf
• #7: *** We’re here to discuss a framework to assess metrics in an agile world. <click>External: Money. Praise. Status. Balloons and faerie dust. Trade conferences. Internal: Job satisfaction. Pride in quality craftsmanship. Shared successes. Adherence to personal ethics. Self-improvement. Deep learning.Avoid Theory X/“Rational Self-Interest” thinking.
• #8: Informational: Diagnostic. Must avoid creating a behavioral change.Motivational: Meant to encourages change of behaviors and habits (BAD!)You take child’s temp to measure illness, not to punish/reward. Do you chart it?
• #10: <<click> indicators tell us things are getting better<click> they are not moving along the expected/measured/optimal path at allAt first, true value increases (workers don’t really understand the metrics, so behavior doesn’t change).Shortcuts appear.Competition quietly ensues, shortcuts propagate.The metric starts to reward the shortcuts, not the desired outcome.Austin attributed this to Partial supervision (vs. total, or delegation) results in the LETTER but not the SPIRITExternal motivations obscure and deactivate internal motivations (Pink: they also seem to deactivate creativity!)[Austin, p. 16]
• #13: Ten years ago, I would not have eaten these.
• #14: USDA Tomato RedProfessed Intent: Quality/Common language for commerce over phone lines.Result: Seen as a critique of product.Effect1989 EPA draft report indicated higher levels of dangerous pesticides to preserve color and continuityUnintendedConsequencePoint: SERIOUS consequences!
• #16: What are (or were) the benefits of the metric.What unintended consequence resulted.Possible reasons why it failed.
• #18: Types of metrics: Qualitative, fast, inexpensive, unobtrusive.
• #19: Metrics! Usually requires a device. Quantitative. Precision.(Amount of congressional funding for the KnightRider project: $321,146,715 – jumps to …716?)
• #20: <read>How many millions in the KnightRider project?
• #22: Aka Organizational performanceAustin tells us that real output is often difficult to measure.Putnam and Myers point out (perhaps rightly) that it takes so long to measure value on a software project (as compared to other industries???), that other more immediate measures must be taken. Given support time and the whole lifecycle, they have a point.
• #23: Team PerformanceIndivperf? - let’s ignore individual perf and evaluation stuff, and just talk about team performance.
• #24: You know it when you see it but when you try to define it you lose it (Zen and the Art of Motorcycle Maintenance)
• #27: Solicit Metrics/Add to FrameworkCollect themNow our list…
• #28: SLOC is baaaaaad.
• #29: Size? Or Technical debt!
• #30: NOT for team performance.Rob: “It’s just a planning tool!” (Diagnostic).Jim: “It’s cost-oriented. You want to minimize cost (don’t you?)”
• #31: expected-value/time? (business-value points / iteration)
• #32: Cycle time – time from concept to cash (how to measure, e.g. for a story on an agile project?)
• #34: Diagnostic, not performance
• #35: niksilver.com
• #36: AustinAffirming the Consequent: (A => B ) => ( B => A )
• #37: Don’t use external rewards.Turn “motivational” into informational.Take the FUN out of dysFUNctional (Pollyanna Pixton)
• #38: 5 dysfunctions of the team.Avoid individual measures. Measure Up one level.Anonymize data and report on team.And group of teams rather than team.Ultimately, Measure the org.
• #39: Five Core Metrics:The Intelligence Behind Successful Software Management Lawrence H. Putnam & Ware Myers Dorset House Publishing Company, May 2003