A client report executive summary works when it does three things in about three sentences: say what happened, say why, say what comes next. Written that way, from the finished dashboard, it takes maybe eight minutes. Written the way most of them are written, it takes thirty seconds and says nothing, because it was typed before anyone knew how the month ended.
That is the actual mechanism behind every vague summary box. The field sits at the top of the template, so it gets filled first. At the point an account manager is building the report, the export may be loaded but the pacing check is not done, the conversion numbers are still settling, and nobody has looked at week four against week one. So you write something that cannot be wrong yet. “Overall, performance was strong this month.” Then you go do the real work, finish the charts, scan the top of the page, see a sentence, and assume it was reviewed.
Flipping the order fixes most of it. Build the summary last, pull every claim in it from a tile or chart that is already on the page, and let the dashboard do the arguing underneath.
The only part of the report some clients read in full
A CMO who gets four agency reports a month is not opening all four dashboards. A brand manager forwarding your report to a finance lead will paste the summary into the email and attach the link. People who do open the report read top to bottom for about fifteen seconds before deciding whether to keep going.
Which means the summary is not a courtesy introduction to the report. It is often the report. Everything below it is evidence for people who want evidence, and a reference for the quarterly review in three months. The report summary clients read is the one that tells them something they did not already know by Thursday of last week, in language they could repeat to their own boss without checking anything.
That raises the bar on specificity more than on polish. Nobody forwards “performance was strong.” People forward “we hit 94% of the plan and CPA came down 22% after we cut the two worst placements in week two.”
“Performance was strong this month” survives every month unedited
Run a quick audit on yourself. Open the last six reports you sent for one account and read only the summary boxes, in order. If you cannot tell which month is which without looking at the dates, the summaries were never doing a job.
The sentences that survive unedited all share a shape. They describe direction without magnitude (“engagement continued to grow”), they attribute without mechanism (“thanks to ongoing optimisation”), and they promise without commitment (“we will continue to monitor and optimise”). None of those can be checked against the dashboard, which is exactly why they feel safe to write early.
A useful test before you send: could this sentence be false? If there is no version of the month in which the sentence would have been wrong, it is not information.
Write the client report executive summary last, against the finished dashboard
The practical sequence, for a monthly report:
- Load the export and finish the dashboard. Tiles, charts, tables, breakdowns, all of it.
- Do the reconciliation against the media plan. Spend versus budget, impressions versus booked, delivery per line item.
- Read your own dashboard as the client, once, without defending it. Note the first question you would ask.
- Write three sentences that answer: what happened, why, what next.
- Reread the summary next to the top row of tiles and check that every number in the text appears on the page.
Step three is the one people skip, and it is where the summary actually comes from. The question you would ask if you were the client is almost always the thing the summary should pre-empt. If you look at your own report and immediately wonder why CPM climbed in the last ten days, write that, because they will wonder the same and they will email you about it. That is also the difference between a report that closes a loop and one that starts a thread, which is worth thinking about alongside how you order the sections in the first place.
Doing this against a static deck is harder than it sounds, because the deck gets locked and exported days before the client opens it, so the summary is written against numbers that may still move. Reporting on a live link removes that gap: the dashboard the client opens is the dashboard you wrote the narrative from. Screebl works this way by design, which also means the summary text sits in the same reusable template as the layout, as a text block you refill per report rather than a paragraph you copy forward from last month.
Three sentences, three jobs
What happened. One claim with a number and a reference point. Not “we delivered 2.4 million impressions,” which is a fact without a verdict, but “we delivered 2.4 million impressions against 2.3 million booked, finishing the flight on plan.” The reference point can be the media plan, the previous period, the target, or the benchmark you agreed at kickoff. Pick one and stay with it across months so the client learns to read the comparison.
Why. The mechanism, in the agency’s own words, naming what you did or what the market did. “The improvement came from shifting 30% of display budget into the retargeting segment from week two.” This is the sentence clients cannot write themselves, and the one that justifies the retainer. It is also where people hide behind “optimisation,” a word that means nothing to the person paying for it.
What next. One commitment with a timeframe. “Next month we are testing two new video cuts against the top-performing audience and will report CPV by cut.” If there is nothing to commit to, say what you are holding steady and why.
A fourth sentence is fine when there is a decision waiting on the client, like a creative approval or a budget shift. More than that and you are writing the report twice.
Lead with the number that mattered this month
The headline tile on your template and the headline number in your summary do not have to match. The tile layout is fixed because consistency across months is worth something. The summary is where you get to say which number carried the story this time.
Suppose a display flight finished at a 0.09% CTR, in line with every other month. The CTR tile is at the top of the dashboard because it always is. But this month the real event was that a frequency cap change pushed average frequency from 7.2 to 4.1 and reach up 38% on the same spend. That is the lead, even though frequency is three tiles down and CTR is the one the client’s eye lands on first. Choosing between those is really a question about which metrics carry meaning for the objective you were briefed on, and the summary is the one place in the report where you make that choice explicitly instead of letting the layout make it for you.
Be careful with the inverse move. If the number that mattered is bad, leading with a good secondary metric reads as evasion to anyone who scrolls.
When the month went badly
The three-sentence structure holds, but the weight shifts. What happened gets the plainest sentence in the report, with the number in it. Why gets more room than usual, because a cause the client can understand is what separates a bad month from an agency that has lost control of the account. What next gets the most specific commitment you have ever written, with a date.
What does not work in a bad month: softening the first sentence so the client has to read to the end to find out something went wrong, or burying the number in the body and summarising it as “a challenging period.” People who feel managed stop reading summaries altogether, and then you have lost the one part of the report you knew they opened. There is more on handling the whole conversation in reporting bad results without losing the client, but the summary itself is where the tone gets set, before anyone has seen a chart.
One more habit worth building: keep the old summaries. Six months of three-sentence narratives, each with a commitment and a result, is the clearest account history you will ever have, and it is considerably more persuasive at renewal than a folder of PDFs. If you want to see how the live-link version of that looks in practice, the plans differ mainly in whose brand sits on the report.
