Skip to content

1 or 2 credits, some free · 1 credit is about $1

pocketshelf

Pro PM Prompt Pack: Data and Decisions

50 advanced prompts plus a Claude Project / Custom GPT setup file for turning data into decisions.

ForProduct managers

  1. 1.Frame an analysis question

    Turn this vague question into a precise analysis plan: [VAGUE_QUESTION]. Give the metric, segments, time window, data needed and how the answer changes a decision.

  2. 2.Build a simple forecast

    Build a simple range forecast for [METRIC_NAME] over [PERIOD] from this history: [HISTORY]. State assumptions and give best, expected and worst cases.

  3. 3.Design a dashboard

    Design a one-screen dashboard for [TEAM_NAME]: the five metrics, why each matters, how to segment it, and the alert threshold.

  4. 4.Explain statistics to executives

    Explain [CONCEPT] to executives in 100 words with one example from [INDUSTRY].

  5. 5.Find leading indicators

    Suggest leading indicators for [LAGGING_METRIC] and how to test whether each really predicts it.

  6. 6.Estimate the size of an opportunity

    Estimate the size of the opportunity for [IDEA] with a bottom-up model. List every assumption and rate my confidence in each.

  7. 7.Decide with incomplete data

    We must decide [DECISION] with incomplete data: [AVAILABLE_DATA]. Give a decision rule, what would change it, and a small step that buys information.

  8. 8.Write an insight summary

    Write a 5-bullet insight summary from this analysis for people who will only read the top: [ANALYSIS_NOTES].

  9. 9.Check a result for a data error

    This number looks surprising: [SURPRISING_NUMBER]. List the ways the data could be wrong (logging, joins, duplicates, time zones, filters) and a quick check for each before we act.

  10. 10.Choose the right average

    For [METRIC_NAME] with this distribution: [DISTRIBUTION_NOTES], say whether the mean, median or a percentile is the honest summary and why.

  11. 11.Spot Simpson's paradox

    These overall and segment numbers seem to disagree: [OVERALL_AND_SEGMENT_DATA]. Say whether a mix shift explains it and which view should drive the decision.

  12. 12.Judge whether a sample is big enough

    We have [SAMPLE_SIZE] users and saw a difference of [OBSERVED_DIFFERENCE]. Explain in plain words how sure we can be, and how many more we need to be confident.

  13. 13.Explain correlation versus cause

    We found that users who do [BEHAVIOR] retain better. List the ways this could be misleading and the cheapest experiment that would show cause.

37 more prompts in the full pack

What's inside

Format
Markdown, JSON, TXT
Contents
50 prompts
Version
1.0 · Updated 3 Oct 2026
Licence
Personal What this means
Checked
Checked before it ships
  • pro-pm-prompt-pack-data-and-decisions.mdMarkdown · 9 KB
  • pro-pm-prompt-pack-data-and-decisions-project-setup.jsonJSON · 3 KB
  • pro-pm-prompt-pack-data-and-decisions.txtTXT · 9 KB
  • LICENSE.txtadded to every download

About this item

A pro pack for senior product managers and founders who need more than a template. It holds 50 distinct prompts for turning data into decisions: analysis plans, metric definitions, forecasts, experiment readouts, bias checks and explaining numbers to executives. Each prompt sets up a role, the inputs you must supply and the exact shape of the answer, and is written for the pack rather than adapted from another one. The pack also includes a setup file (pro-pm-prompt-pack-data-and-decisions-project-setup.json) with instructions and conversation starters for a Claude Project or Custom GPT. You get the prompts as Markdown and plain text so you can paste them anywhere. Placeholders are clearly marked in [UPPER_CASE].

Reviews

No reviews yet.

You might also grab