How to Write an Executive Summary from Data (with Examples)

An executive summary from data tells a busy reader, in under a minute, what happened, why it matters and what should be done. Lead with the single most important finding, back it with two or three numbers that include a comparison, and end with a specific recommendation. Everything else belongs in the body of the report.

The four-part formula

Almost every good data summary follows the same shape. Call it What, Why, So what, Now what:

1. What happened: the headline result, with a number and a comparison. ‘Revenue for August was R1.26 million, 7% below July.’
2. Why: the main driver you can see in the data. ‘The drop came almost entirely from the retail channel; online sales were flat.’
3. So what: the implication, risk or opportunity. ‘At this rate we will miss the quarterly target by about R300,000.’
4. Now what: the recommended action, ideally with an owner. ‘Sales to review retail pricing with the three largest stores before 15 September.’

That’s four sentences. Many excellent executive summaries are not much longer.

Before and after: three examples

Example 1: Monthly sales

The ‘before’ version uses words but says nothing. The ‘after’ version could be acted on by someone who reads nothing else.

Example 2: Customer support

Example 3: Operations

All numbers above are illustrative, but notice the pattern: every figure has a comparison, every claim points to a cause, and the last sentence is an action.

How to find the headline in your data

The hardest part is deciding what matters most. Work through these questions in order:

• What changed most compared with last period or target? Rank your KPIs by size of change.
• Where is the change concentrated? Break the biggest mover down by region, product, channel or customer. Often one segment explains most of it.
• Is it a trend or a blip? Check three to six periods, not just two.
• What would the reader do differently if they knew this? If nothing, it probably isn’t the headline.

A quick way to find concentration: if total sales fell by R95,000 and one region fell by R88,000, that region is your story. Our guide on finding trends in sales data covers more techniques.

Common mistakes

Using AI to draft the summary

Language models are good at writing clear sentences and poor at arithmetic. The safe way to use them is to calculate the figures first, with a spreadsheet or code, and then ask the AI to write the summary using only those figures. Never ask a chatbot to read raw data and ‘summarise the key numbers’ without checking every figure it produces. We explain why in how to stop AI making up numbers.

Summarix works this way by design: it computes KPIs, comparisons and charts from your uploaded file or connected database with its own code, then the AI writes the executive summary, insights and recommendations around those figures. You get a first draft in about a minute and can edit it or ask follow-up questions before you share it.

A reusable checklist

• Is the first sentence the most important finding?
• Does every number have a comparison?
• Have you explained the main driver?
• Is the implication clear (risk, cost or opportunity)?
• Is there a specific action, ideally with an owner and date?
• Can it be read in under a minute?
• Would it still make sense if it were the only page someone read?

Write the summary last, after the analysis, but put it first. It’s the part of the report most people will read, and often the only part. For turning findings into a narrative people act on, see data storytelling for business reports.