The most common failure in analytics isn't wrong analysis. It's right analysis that never changes anything. Someone spends a week digging into returns, finds something real, writes it up as forty slides of charts, and the meeting ends with "interesting, let's keep an eye on it."
How you write up the work matters as much as the work itself. Here's the structure I use, and why each part is there.
Key takeaways
- Lead with the recommendation, not the method.
- Show the two or three pieces of evidence that carry the conclusion, and move the rest to an appendix.
- Say how confident you are, and what would change your mind.
- End with a decision someone can make in the meeting, with an owner and a date.
The structure
Start with the answer. The first paragraph says what you recommend and why, in plain words. "We should stop offering free returns on sale items, because they account for most of our return costs and very little of our margin." A busy reader who stops there still has the point.
Then the evidence, but only the evidence that matters. Usually two or three findings carry the conclusion. Show those clearly, one chart each, with a sentence saying what the chart shows. Everything else you looked at goes in an appendix for anyone who wants to check your work.
Then the honesty section. How confident are you? What data was missing? What would make you change your mind? This part builds more trust than any amount of polish, and it stops the conversation from turning into an attack on the method.
End with the decision. One clear ask: approve this, test that, or stop doing something. With a name and a date beside it.
Writing so it travels
Your write-up will be forwarded to people who weren't in the meeting. A few habits make sure it still works when you're not there to explain it:
1
Title every chart with its conclusion
"Sale items drive most return costs" instead of "Returns by category". The reader shouldn't have to work out what the chart is saying.
2
Define every metric once
If "return rate" means returned orders divided by all orders, say so the first time it appears.
3
Keep the numbers checkable
Note the date range and the source for each figure, so anyone can reproduce it later.
A test worth applying
If the recommendation would still be clear after someone read only the first paragraph and the chart titles, the write-up is doing its job.
The point of all this
Analysis exists to change what a business does. Writing it up so the decision is easy is the last, and most often skipped, part of the job. Get it right and the next request for analysis comes with a question attached, which is where the most useful work starts.
