SEO report template

Most SEO reports lead with traffic and a table of keyword rankings. Both are the wrong things to lead with, and the second one is close to indefensible.

This template leads with conversions from non-paid search, and reports grouped query performance instead of individual rankings.

Preview of the file

Why there are no keyword rankings in it

This is the part people push back on, so here is the reasoning.

An average position in Search Console is your own measured data across everyone who saw you. A rank-tracker number is a single check from one location on one device — and results are personalised, so two people searching the same term at the same moment see different pages. Reporting that number as a fact invites a conversation about why “we dropped to position 4”, which is often not a real event at all.

What goes in the report

Section	What it contains	Why it is there
Headline	One sentence: did non-paid search get better or worse this month, and by how much.	Most readers stop here. If they cannot tell from this line, the report has failed regardless of what follows.
Non-paid conversions	Conversions and enquiries from non-paid search, against the previous period and the same month last year.	The only number the business genuinely cares about. It goes above traffic deliberately.
Traffic and visibility	Clicks, impressions, average position and click-through rate from Search Console.	Search Console is your own data rather than a tool estimate, which is why it is the source here.
Query groups	Grouped queries — branded, commercial, informational — with clicks and position for each.	A total hides the story. Branded growth and commercial decline can average out to "flat".
Page performance	Pages gaining, pages losing, and pages that earned their first clicks this period.	First-click pages are the earliest evidence new work is landing, usually weeks before totals move.
Technical health	Indexed vs submitted, crawl errors, Core Web Vitals status, and anything newly broken.	A technical problem is the one thing that can undo months of work between two reports.
Links earned	New referring domains worth naming, and anything lost.	Counts alone are noise. Three relevant domains beat forty scraped ones, and only naming them shows that.
Commentary	What changed, why, and whether it was intentional.	The part a blank template cannot give you, and the reason anyone keeps reading your reports.
Next actions	Three at most, each with an owner and a date.	A report with no action is a status update. Three is the most anyone acts on.

The order is deliberate. Conversions sit above traffic because nobody buys a session, and the headline sits above everything because most readers stop there.

What each column is for

Column	What to put in it
Month	One row per month. Twelve rows makes a trend readable instead of a snapshot.
Clicks	Search Console clicks, non-paid only. Not sessions — they count differently.
Impressions	From Search Console. Rising impressions with flat clicks means visibility without relevance.
CTR	Calculated. Falling CTR at a stable position usually means the result no longer matches the query.
Avg position	From Search Console. An average across queries, so treat a small move as noise.
Conversions	From non-paid search specifically. Set this up before you need it.
Conv. rate	Calculated. The check on whether extra traffic is the right traffic.
Ref. domains	New referring domains this month. Name them in the report, do not just count them.
Indexed pages	From the Search Console Pages report. A sudden fall is the most urgent thing on this sheet.

Two of those are worth expanding. Impressions rising while clicks stay flat means you have gained visibility for queries you do not deserve to win — normal when a site grows, and not a problem unless CTR is falling on the queries that matter. And indexed pages falling suddenly is the most urgent line on the sheet: it is usually a technical change rather than an algorithm, and it is fixable within days if you notice.

The completed example

The PDF is one month for an invented business, written the way a report should be written. Two things in it are worth copying.

The query group table shows informational traffic as the largest group producing one enquiry, while commercial queries at under a quarter of the clicks produced twenty-six. That single table changes which work gets funded next quarter.

And the commentary states a decline and explains it — CTR fell year on year despite a better average position, because the site now appears for a wider set of queries. A decline you explained is a credibility gain. A decline the client finds themselves is how engagements end.

Mistakes, and what to do instead

Mistake	What to do instead
Reporting keyword rankings for single terms	Report grouped query performance from Search Console. Rankings differ by person, device and location, so a rank-tracker number is one sample presented as a fact.
Leading with traffic	Lead with conversions from non-paid search. Traffic is an input; nobody buys a session.
No comparison period	Compare with the previous period AND the same month last year. Seasonal businesses look broken every January without the year-on-year line.
Counting links instead of naming them	Name the referring domains worth having. Forty scraped directories and three relevant sites are not the same result.
Sending a tool export	A dashboard screenshot is not a report. If nobody wrote a sentence about what it means, nothing has been reported.
Hiding a decline	State it in the headline with the reason. A decline you explained is a credibility gain; a decline they find themselves is the end of the engagement.

Using it in Google Sheets

File → Import, choose the workbook, and select “Replace spreadsheet”. Every formula here is standard — IFERROR, SUM, LOOKUP — so nothing breaks in the conversion. The yellow input shading survives; only the data validation dropdowns occasionally need reapplying.

If you would rather report every channel rather than search alone, the marketing report template is the same idea across paid, email and social, and the two are designed to sit alongside each other.