Audience Depth Guide
How to calibrate methodology explanations for different reader types. Match the depth to who will read it, not to how proud you are of the method.
Tier 1: Executive / Decision-Maker
Who: C-suite, VP, Board, senior business leaders
What they care about: Can I trust this? What's the conclusion? What should I do?
What they don't want: Methodological detail, statistical terms, caveats that don't change the decision
Write this much: 1 paragraph maximum
Include: Data source, time period, method in one plain sentence, single most important limitation
Omit: Formula, model parameters, sample sizes (unless very small), confidence intervals (unless the decision depends on them)
Test: Could a smart 10-year-old follow the logic? If yes, it's right for this audience.
Tier 2: Business Analyst / Domain Expert
Who: Product managers, finance analysts, operations leads, marketing strategists
What they care about: Is the method appropriate? Did we account for [known issue]? Can I replicate or extend this?
What they don't want: Dense statistical notation, code, heavy academic framing
Write this much: 3–5 paragraphs or 1 structured writeup section
Include: Data sources and joins, time period and filters, method logic in plain English, assumptions, key limitations
Omit: Code, mathematical notation, model internals beyond "how it works at a high level"
Test: Could a product manager on the team follow the logic and explain it to their skip-level?
Tier 3: Technical Peer / Data Scientist
Who: Other analysts, data scientists, data engineers reviewing the work
What they care about: Correctness, reproducibility, statistical validity, code quality
What they don't want: Oversimplification that hides important choices
Write this much: Full technical appendix or inline code documentation
Include: Exact SQL or code, model hyperparameters, feature engineering choices, train/test split details, validation approach, assumptions with quantified sensitivity
Omit: Nothing — this is the audience that needs the full picture
Common Calibration Mistakes
| Mistake | Consequence | Fix |
|---|---|---|
| Using Tier 3 language for Tier 1 audience | Executive stops reading; decision delayed | Rewrite the summary section using the translation table |
| Using Tier 1 depth for Tier 2 audience | Business analyst can't validate; asks repeated questions | Add a data + method section |
| Writing only for Tier 1 and having Tier 3 in the same doc | Technical reviewers can't verify; executives get overwhelmed by appendix | Use layered structure: summary → detail → appendix |
| No limitations section for any tier | Stakeholder over-trusts the output | Every methodology explanation needs at least one limitation |
Quick Calibration Checklist
Before finalising a methodology explanation:
- I know who will read this (name the primary audience tier)
- The depth matches that tier
- Statistical terms are translated for Tier 1 and Tier 2 audiences
- At least one limitation is explicitly stated
- The "how we got here" is short enough that it doesn't compete with "what we found"