From problem to decision
1. Map the baseline
Choose one period and segment. Count enquiries, qualified leads, offers and purchases. Calculate each transition as the next stage divided by the previous stage × 100%. Do not compare an unfinished cohort with a mature one.
Period, segment, counts and conversion rates. Which data is still missing?
2. Choose one cause
A large drop is not always a defect: some leads are not a fit. Examine several lost deals and separate observations from explanations. Compare the advertising promise with the buyer’s actual decision criteria.
Where is the loss? What evidence supports this? What alternative explanation is possible?
3. Design the test
Change one element: segment, value proposition or a sales action. Set the main metric, duration, cost limit and stopping rule. With few deals, results are a signal rather than causal proof.
Change → metric → baseline → target → duration → owner → guardrail.
4. Record the decision
After the test, record actual results, evidence limitations and a decision: continue, revise or stop. Leave this step incomplete until the test has finished.
What happened to the metric? What remains uncertain? Decision and next action.
How to judge whether it helped
You should leave with a document that supports a decision and a next action. Completing steps does not guarantee a commercial result. Measure the chosen change separately over an agreed period. If evidence is missing, state that and plan how to collect it.
How this differs from S01–S13
This is a user route for a specific problem. The GRI AI research stages S01–S13 and the 13-week programme remain in the library as separate methodological work. This route is not a validated GRI diagnostic scale.