Run a data analysis from raw numbers to a decision-ready recommendation
Most analysis stops at "here's what the data shows," short of what to actually do about it.
- Outcome
- A specific finding, checked against alternative explanations, translated into a recommendation someone could actually act on.
- Time
- 45-90 minutes depending on data complexity
This workflow is part of PromptKit Pro
A handful of prompts and workflows stay free for everyone — Pro unlocks the full library, including this one, plus the guided runner.
Data analysis often stops one step short of useful: it finds a pattern, narrates it, and stops, leaving the translation into "so what should we do" as an unstated exercise for whoever reads the analysis. This playbook runs the full chain — find the real pattern, check it isn't a confound, then translate it into a specific recommendation — so the output is something a decision-maker can actually use.
When not to use this
If you just need the numbers reported accurately for a dashboard, this is more than that task needs — a clean summary is the right output there. This is for analysis meant to inform a real decision.
Did this work?
You end with a specific recommendation tied to a checked finding, not just a narrated summary of what the numbers show.
Tested on claude-opus-5. Evidence status is draft; it moves to battle-tested only on recorded runs, never by hand.