A COMPLETE USE CASE · ABOUT 8 MINUTES
Understand what is working
“I have results, but I don’t know what to change next.”
Twenty sample visits became eight signups and four activations. Some attribution is missing and the cohort is only three days old. The useful next move starts with what those numbers can—and cannot—say.
Teams reviewing proposal outcomes or a product launch.
Recorded outcomes, date windows, known sources and a question to investigate.
An honest evidence review and a measurable next experiment.
What needs to be set up in the real app?
Actual recorded data in the appropriate workspace. Launch/visitor metrics depend on real instrumentation and applicable consent.
Tools along the way: Analytics → Launches → Results → The PM → Human decision
Where does your story start?
Choose the useful part. Earlier sample outputs are supplied, not credited as your work.
Follow the whole fictional story, from the first decision to the final result.
0 of 5 decisions kept on this path. All practice is fictional and local.
THE PM HELPS HERE
Put the right window around the numbers
This sample launch is three days old. You want to report day-7 retention.
A three-day-old sample cohort
20 visits → 8 signups → 4 activations. Day 7 has not occurred.
- All numbers are fictional
THE WHOLE RECIPE, WHENEVER YOU NEED IT
Take the path into Seerist.
Use the same order with your actual facts and permissions. The example outputs here are fictional teaching material.
Put the right window around the numbers
Analytics / Launches → Results
A missing or incomplete observation is not a zero. Compare like-aged, fully observed windows.
Open the reference lesson →Keep missing sources visible
Analytics → attribution
Preserve real references through signup. First-party visitor signals do not identify anonymous people or infer their company.
Open the reference lesson →Separate a clue from a conclusion
AI Employees → PM review
The PM can explain uncertainty and propose what to measure. Recommendation is distinct from authority to execute changes.
Open the reference lesson →Choose a metric before making a change
Launch review / PM recommendation
Save the starting snapshot, target metric and observation window. Ask a human to choose the experiment before execution.
Open the reference lesson →Make the recommendation accountable
Saved review → metric recheck
A later comparison can support another decision. It should report small samples, incomplete attribution and uncertainty instead of inventing uplift.
Open the reference lesson →
Source-based guidance checked against Seerist app a365066 · September 13, 2026. Each linked reference lesson identifies its product sources. Guide decisions are learning exercises, not proof of live provider behavior or business outcomes.