Audience Depth Guide for Analysis Documentation
Calibrating documentation depth
The same analysis can be documented at three different levels depending on who will read it. Over-documenting wastes time; under-documenting creates ambiguity that surfaces as repeated questions.
Level 1 — Executive / Stakeholder
Who reads this: VP, Director, non-technical business owner.
What they need:
- The answer, not the method
- Confidence that the analysis is trustworthy
- What decision this enables, and what it does not
Format guidance:
- Max 1 page (or equivalent)
- Lead with the key finding, follow with supporting evidence
- One chart maximum; no SQL, no technical terminology
- Explicit recommendation or next step
Avoid:
- Methodological detail ("we used a z-test at alpha=0.05")
- Data caveats beyond the most critical one
- Intermediate steps or alternative approaches considered
Level 2 — Analytical Peer / Manager
Who reads this: Senior analyst, analytics manager, product manager with data literacy.
What they need:
- Methodology at a level that lets them evaluate whether it is sound
- Key assumptions and their confidence level
- How to reproduce or extend the analysis
Format guidance:
- 2–4 pages
- Context → Approach → Findings → Caveats → Next steps
- Include the main SQL or code references (not the full script)
- Explicit assumption log with confidence ratings
Avoid:
- Full code dumps inline
- Raw table dumps; use summarised outputs
- Over-explaining standard techniques
Level 3 — Future Analyst / Handover
Who reads this: Someone who will extend, debug, or re-run this analysis in 6+ months.
What they need:
- Complete reproducibility: where data lives, how to run the code, expected outputs
- All design decisions and alternatives considered
- Known issues and workarounds
Format guidance:
- No page limit
- Every query, script, and file path referenced
- Step-by-step instructions to reproduce
- Change log if the analysis is run repeatedly
Avoid:
- Assuming context that is obvious now but won't be later
- Skipping the "why" behind non-obvious decisions
Choosing the right level
| Trigger | Level |
|---|---|
| One-time ad-hoc for a meeting | 1 |
| Recurring report or dashboard | 2 |
| Strategic decision with board visibility | 1 + summary of 2 |
| Handover to another team | 3 |
| Reusable analysis template | 3 |
| Internal methodology reference | 2–3 |
Reusable sections across levels
These sections appear in all levels, just at different depths:
- Question — what business question this answers
- Data sources — tables, systems, date ranges used
- Key finding — the headline answer in one sentence
- Caveats — what could be wrong or missing
- Next steps — what action follows from this