Turn analysis into a decision-ready story

One-page decision memo from a churn analysis

Take the Telco customer churn dataset, run the analysis properly, then throw away everything except one page. The deliverable is an executive memo that opens with a recommendation, backs it with two or three charts whose titles state the takeaway, admits one honest limitation in the data, and proposes the next experiment. The discipline is subtraction: the notebook can be long, the memo cannot.

Telco Customer Churn dataset on Kaggle — 7,043 customers with contract, tenure, billing and churn label

1. Run the churn analysis and note every finding that looks decision-relevant · ~2.5h
2. Cut to the two or three charts that carry the recommendation and retitle them as takeaways · ~2.5h
3. Write the one-page memo: recommendation first, caveat, proposed next experiment · ~2h
4. Get one person to read it cold, then fix whatever they misread · ~1h

• Memo fits on one page/slide and leads with the recommendation in the first sentence
• Exactly 2-3 supporting charts, each with a title that states the takeaway, not just the metric name
• Includes one explicit caveat about a confound or data limitation
• A different person (peer review) can read it in under 2 minutes and repeat back the recommendation correctly

Evidence a recruiter can check

• The one-page memo itself, opening with the recommendation in the first sentence and naming the caveat out loud
• The two or three charts with takeaway titles, shown next to the raw default-styled versions they replaced
• The full analysis notebook behind the memo, so a reader can check that the number in the recommendation is real
• A note of what the cold reader repeated back after two minutes, and what you changed because of it

Turned a churn analysis into a one-page decision memo — a single recommendation, three takeaway-titled charts, a stated data caveat and a proposed follow-up experiment, tested on a cold reader before sending.