The executive summary is the paragraph at the top of a monthly marketing report. It is often the only part a busy client reads before they decide whether to open the rest. It should answer three questions in under a minute: did last month work, why, and what happens next.
Generate your first client report free — upload a Google Ads, Meta Ads, or GA4 CSV, get a scorecard computed in code, and edit a drafted executive summary, wins, concerns, and next-month recommendations before you share.
What an executive summary is (and is not)
A summary states outcomes. A recap lists activity. "We refreshed four ad groups and added negative keywords" is a recap. "Leads rose from 40 to 51 while spend held flat" is a summary. Put the activity further down the report, where it can explain the numbers.
The summary is also not the place to define metrics, list every channel, or make a case for more budget without a figure behind it. If a sentence would still make sense with no numbers in it, it probably belongs somewhere else.
Before you write: a five-minute checklist
Do these against the export, not from memory:
- Name the period and the comparison period (for example, October versus September). Make sure both use the same date length.
- Pick the one metric the client cares about most: leads, purchases, revenue, CPA, ROAS, or sessions to a key page. Lead with it.
- Total spend, conversions, and the efficiency metric for both periods, and calculate the change. Check the arithmetic once.
- Confirm nothing changed in tracking: same conversion action, same attribution setting, no broken tags. If something did change, that goes in the summary.
- Write down the one decision you want from the client, or "none this month".
If you only have one month of data, there is no change to report. Say that it is a baseline month instead of comparing to nothing.
A five-line structure
Most good summaries fit this order. Each line is one or two sentences.
- Headline result: the main metric, this period versus last, with the change.
- Driver: the change you made or the event that best explains it, if the data supports one. Label a guess as a guess.
- Concern: the one thing that did not work or needs watching. Leaving it out makes the rest less believable.
- Next step: what you will do next month, tied to the metric in line one.
- Ask: what you need from the client, or say nothing is needed.
Keep the whole thing to roughly 60–120 words. If it runs longer, move the detail into the "what changed" section of the report.
Example executive summaries
All examples below are illustrative. The clients, figures, and causes are invented to show the structure. They are not benchmarks or real results. Replace every number with the client's own export.
Example: a strong month (Google Ads, lead generation)
Example: "October Google Ads produced 51 leads, up from 40 in September (+28%), on almost the same spend ($3,050 versus $3,000). Cost per lead fell from $75 to $60. Most of the gain came from the two service campaigns where we moved budget mid-month. The concern: the brand campaign's cost per click rose 15%, which we are watching for competitor bidding. In November we will extend the same budget shift to the third service line and keep total spend flat. Nothing needed from you this month."
Example: a flat month (Meta Ads, ecommerce)
Example: "October Meta Ads drove 212 purchases versus 208 in September, with spend flat at $6,400 and ROAS steady at 2.4x. Flat is acceptable here because we held budget while the new product line launched on the site. The concern: the two oldest ad creatives now take 40% of spend with falling click-through rates. Next month we will rotate in the three new creatives you approved and judge them on cost per purchase after two weeks. Please confirm the November promo dates by Friday so ads and the site match."
Example: a bad month (Google Ads, ecommerce)
Example: "October Google Ads revenue fell to $18,200 from $24,900 (−27%) while spend held at $5,600, so ROAS dropped from 4.4x to 3.3x. Finding: the shopping campaign lost most of its volume after 14 products were disapproved in the feed on October 9. We fixed the feed on October 21, and daily revenue has returned close to the September average since then. Next month we will add a weekly feed check so a disapproval is caught within days. No budget change requested."
Example: a mixed signal (GA4 plus ads)
Example: "October sessions from paid channels rose 18% versus September, but form submissions in GA4 fell from 96 to 81 (−16%). Ad platform conversions did not fall by the same amount, so the gap may be in tracking rather than demand. Hypothesis: the contact form update on October 12 changed the thank-you page that GA4 counts as a conversion. We have asked your developer to confirm. Until that is checked, we are not changing bids. Decision needed: approve a one-hour tracking audit."
Example: a first month (baseline)
Example: "This is our first full month, so there is no prior period to compare. October Google Ads spent $2,400 and recorded 22 conversions at $109 each. This becomes the baseline for future reports. Conversion tracking is set up on the quote form and phone clicks, and we confirmed both fired correctly in a test on October 3. In November we will pause the three keywords that spent over $150 without a conversion and report the change against this baseline. Nothing needed from you."
Phrases that help, phrases to cut
Useful patterns:
- "[Metric] was [this period] versus [last period] ([change])."
- "Finding: …" for things the export or a dated change proves.
- "Hypothesis: …" or "Working theory: …" for causes you have not proven.
- "Next month we will … and judge it on [metric]."
- "Nothing needed from you this month."
Cut or rewrite:
- Adjectives with no number ("strong", "solid", "exciting") unless a figure follows them.
- "The algorithm" as an explanation by itself.
- A result that is not in the export, or a percentage you did not calculate from it.
- More than one ask. If there are three decisions, the summary names the most important one and the rest go in recommendations.
Where ClientNarrate fits
ClientNarrate computes the scorecard (spend, clicks, CTR, CPC, conversions, CPA, ROAS, sessions, and period-over-period change) in code from your CSV export. Then it drafts the executive summary and the rest of the commentary from those figures only. You edit the draft before a client sees it. It does not log into ad accounts, so the export you upload is the source of every number.
Related guides: How to write monthly client report commentary covers the sections below the summary, and How to explain a drop in performance to a client goes deeper on bad months.