Lesson 27: Scale the Playbook to a Second Flow – Copy, Don’t Reinvent
The first flow you fix is a project. The second one is where you find out whether you built a method or just got lucky.
And there’s a specific temptation at this point that I want to name, because I fell for it. Having watched the loop work once, I wanted to improve it before running it again. Better templates. A tighter sequence. Maybe a tool.
Don’t. Run it again exactly as it was. You learn far more from the second identical run than from a first run of something new.
The Rule: Repeatability Beats Novelty
A playbook that gets run once is a story you tell at conferences. A playbook run three times is an operating capability, and the difference is worth more than any refinement you could make between runs.
Keep the cadence identical: visibility sprint, single signal, small governance artifact, pilot, retro. Reuse the templates as they are – the process map, the one-page brief, the SOP format. Keep the experiment window the same length.
Change exactly one thing. The owner.
A Real Example: The Second Flow Beat the First
We had a billing flow that worked. Genuinely worked – the numbers held past the point where I’d have expected them to drift. Onboarding, two teams over, was still struggling with exactly the problems billing had solved four months earlier.
The obvious move would have been for me to run onboarding too. I didn’t, and that was the best decision I made that quarter.
We handed the playbook to a new owner in the onboarding team. Same five-step cadence. Same templates, unedited. Same sixty-day experiment window. My role was to answer questions, not to run it.
The second flow improved faster than the first. Time to value shortened by around eighteen percent inside sixty days, where billing had taken closer to ninety to show comparable movement.
The reason wasn’t that onboarding was easier. It was that the arguments were already settled. Nobody had to debate whether a visibility sprint was worth ten days, because billing had already answered that. The templates existed. Expectations existed. The new owner spent her time on the actual problem instead of on convincing people to let her try.
How To Do This Yourself
Pick your second flow and hold this line:
- Choose a flow with a different owner and a genuinely different failure mode. Two flows that break the same way will teach you one lesson twice.
- Hand over the templates untouched. Every edit you make before the second run is an untested assumption.
- Keep the same experiment window. If you change the duration you can’t compare the runs, and comparison is the entire point.
- Have the first owner spend one hour with the second owner. Not a handover document – a conversation about what surprised them.
- Track both flows side by side and watch for signal divergence. Where the second flow’s numbers behave differently is where your local nuance lives.
Then, and only then, update the templates – using what both runs taught you rather than what one run made you feel.
The Honest Lesson
Scaling exposes nuance, and I handled that badly the first time.
When the onboarding team said “this step doesn’t apply to us,” my instinct was that they were resisting. Sometimes teams are. But often they’re right, and if you insist on an exact copy you’ll get compliance without belief – which lasts about six weeks.
What worked was setting the boundary at the level of the cadence rather than the detail. The five steps are fixed. The experiment window is fixed. Inside that, adapt whatever you need. Teams accept constraints they can move within, and reject constraints that treat them as interchangeable.
The other thing I’d say plainly: resist running the second flow yourself. It’ll go faster, and you’ll learn nothing about whether the playbook works without you in the room. That question is the one that actually matters, because a method that requires you isn’t a method. It’s just you.
Name your second flow this week, and name someone other than yourself to run it.
Next time, I’ll dig into “Run a 30-Minute Metrics Review: Signal, Incidents, Adherence” – see you there.