Choose the problem
Set the baseline, target, metric and guardrails before work starts.
PM + Data + business owner
Output: one-page value brief
5 The two-week value cycle
Review value every two weeks. Change the plan when evidence changes.
A repeatable loop from a measurable problem to a decision about the next use of platform capacity.
One cycle · four steps
Make the change
Agree scope, owners and dependency dates; build a usable slice.
PM + Engineering + QA + dependent teams
Output: releasable change + dependency owner
Read the signal
With tracking ready before release, compare results in each brand to the baseline.
Data + PM
Output: outcome, leading signal and guardrail view
Use the evidence
Review impact and uncertainty; allocate the next cycle accordingly.
PMs + product lead + delivery leads
Output: decision and next owner
Decision
Extend what works · Improve what is inconclusive · Stop what does not create value
The decision becomes an input to the next cycle
What counts as done?
Delivery: working, safe and instrumented. Value: the outcome improves within guardrails. A release alone is not value.
What if the result is weak?
Check adoption, data and dependencies. Adjust or stop; record what changed in the next priority.
The review happens every two weeks; larger changes may span several cycles. Report early signals separately from proven GGR impact.