An SEO report has one job: let the person paying decide what happens next. Most reports fail that test. They are long, they are full of tool scores, and they never state what changed, over what period, compared to what.
Every number names its source and its window
This is the rule I hold myself to. "Organic traffic up" is not a result. "Organic visitors up 112.5% over the same 28 day window, measured in Google Site Kit" is a result, because you can check it and you can repeat the measurement.
On the Klimastål work the two headline numbers were a 30.7% rise in search impressions over a 28 day window in September 2023, and organic visitors up 112.5% across that same window. Both statements carry the window and the tool. If either is missing, the number can mean almost anything.
What goes in the report
Four things, in this order.
1. Visibility. Impressions and clicks from Search Console, with the date range stated, plus the queries and pages that moved. Clicks matter more than average position, because position is an average of averages and hides the pages that actually changed.
2. Indexation. How many URLs are indexed and, when the number moves, the reasons Search Console gives. A drop here is the earliest warning you get that something technical broke, usually weeks before traffic reflects it.
3. Experience measurements. The three Core Web Vitals from field data, against Google's published thresholds of 2.5 seconds, 200 milliseconds and 0.1. Device stated, because mobile and desktop are separate stories.
4. What I did, and what is next. A short list of work completed with dates, and the next few items with the reason each one is next. This is the part clients read first.
What I leave out
Tool scores presented as outcomes. A performance score, a site health percentage or an SEO grade out of 100 is useful for finding work, not for proving it. I will report that a site checkup score reached 96 out of 100 as supporting context, never as the result itself.
Also out: rank tracking screenshots for a handful of flattering keywords, vanity charts with no axis labels, and anything I cannot trace back to a named tool and a date range.
Attribution, stated honestly
SEO reporting sits on correlation. Seasonality, paid campaigns, a press mention or a competitor's own mistake all move the same lines. When I claim a change came from the work, I say what makes me think so: the change is concentrated on the pages I touched, it starts after the deployment date, and it does not appear on comparable untouched pages.
When I cannot show that, I report the change and say the cause is unproven. Clients trust the numbers more, not less, when you are willing to say that.
Cadence
Monthly is enough for most sites. Weekly reporting rewards noise and encourages changing course before anything has had time to be recrawled. What I do check weekly is the indexation report and error counts, because those are alarms rather than trends.
The test for a good report
Read the first page and ask: do I know what improved, by how much, over what period, and what happens next month? If the answer is no, the report is a document, not a decision.
Want this run on your site?
Tell me the domain and the pages that matter. I will come back with what I would fix first and what it should move.
Start a project