6 November 2026

SEO reporting: what your agency should actually show you

An SEO report should tell you what changed, why it matters and what the agency intends to do next. If all you receive is a green arrow beside a ranking, you still do not know whether the work is helping your business.

Useful SEO reporting connects search visibility with the pages being improved, the work actually completed and the enquiries or orders you can verify. It also shows bad news. A report that only includes the winning keywords is a sales presentation, not a reliable account of performance.

This guide shows the reporting detail to ask an agency for, including worked examples of how to interpret it. The numbers below are illustrative, not real client results or claims about Craig's customers. The examples use real reporting measures and ordinary agency deliverables; no private client data has been used.

What should an SEO agency include in a monthly report?

Start with a short summary: the agreed objective, the main changes, the evidence behind them and the decisions needed. Then include a comparable performance view, important landing pages and query groups, verified business outcomes, completed work and the next priorities.

State the dates, data source, country, device and search type where they affect the figures. Compare like with like. October has more days than September; a month-on-month increase can partly reflect that difference. Add a comparable year-on-year view when seasonality matters, and explain any changes to tracking, consent or the website.

Example one: more visibility is not automatically more demand

Imagine a UK service business reviewing the same 28-day period before and after a set of page improvements. Search Console, filtered to UK Web results, records the following:

Impressions: 80,000 previously; 100,000 now — up 25%.
Clicks: 2,400 previously; 2,500 now — up about 4.2%.
Click-through rate: 3% previously; 2.5% now — down 0.5 percentage points.

A weak summary says: “Visibility is up 25%; excellent progress.” A useful summary says: “Exposure increased much faster than clicks. We need to inspect the new queries and landing pages before deciding whether this is commercially relevant growth.”

The agency should show which pages gained impressions, whether queries were branded or non-branded, and whether device mix or search-result appearance changed. Lower CTR does not automatically mean a poor title: broader exposure at lower positions can change the average. A title review might be appropriate, but the report should establish the likely cause first.

Google's Search Console definitions matter here. Impressions measure appearances under its counting rules; clicks measure interactions that lead out of Google under the relevant report rules. CTR is clicks divided by impressions. Average position is an aggregate, not a fixed rank that every visitor saw. Query rows also omit some anonymised queries, so adding visible rows may not reproduce the chart total.

Example two: show qualified enquiries, not just a form event

Suppose GA4 records 42 successful enquiry events associated with Organic Search during a reporting period. The agency should not call those 42 new customers. The owner checks the enquiry records and identifies six spam entries, four recruitment messages and two duplicate enquiries. That leaves 30 suitable enquiries on this illustrative definition.

The report can now show the difference between tracked events, recorded submissions and qualified opportunities. It should define “qualified” with the client, explain how duplicates are handled and identify any missing source information. A click on a telephone number is a click, not proof that a call connected. A contact-button click is not proof that a form reached the inbox.

If eight opportunities remain open, list them as open rather than estimating a win and presenting it as fact. If the business cannot yet connect enquiries with sales, say so. Do not fill the gap with an assumed lead value disguised as revenue.

Ask which GA4 dimension was used: session-level Organic Search and event-level attributed credit are not interchangeable. Consent choices, attribution settings and unrecorded journeys can affect the comparison. Search Console clicks and GA4 sessions also measure different things; they are not expected to match exactly.

For the financial calculation rather than the reporting format, use the separate SEO ROI guide. For key-event and attribution setup, read the GA4 measurement article.

Example three: a ranking improvement needs a page and a purpose

“Twenty keywords reached page one” leaves several questions unanswered. Which keywords? In what country and device? Were they agreed commercially relevant phrases, or searches nobody in your market uses?

A useful entry might read: “The boiler-servicing landing page moved from position 14 to 8 in our scheduled UK mobile rank check for this query. Search Console clicks to that page increased from 90 to 120 across comparable 28-day periods. Two suitable enquiries were recorded, but the sample is too small to claim a stable improvement in lead generation.” This is another worked illustration, not an actual business result.

The agency should supply the query, destination URL, observation date, location and device, plus context from page performance. A third-party rank check is a snapshot. Search Console's average position describes observed search appearances differently. Do not combine the two as if they were one continuous measurement.

Show the work completed, not just the movement afterwards

Clients should be able to see what they paid for. A change log is more useful than a line saying “technical optimisation completed”. Ask for affected URLs, the problem identified, the implementation date, a verification result and anything still waiting on the client or developer.

For example: “Updated the internal links on three service pages to point directly to their final URLs; checked the destination responses; two legacy redirect issues remain with the developer.” That gives the client something to inspect. It does not claim that the work caused all subsequent traffic growth.

For content, list the pages commissioned, edited and published, the question each answers and the next review step. For link work, explain the actual activity and show verifiable placements where relevant. Third-party authority scores and estimated traffic can provide context, but they are not Google's own measures or proof of a valuable link.

Report drops and uncertainties before the client finds them

If a valuable page loses clicks, include it even when overall traffic is up. Explain what is known, what is being tested and what remains uncertain. A sensible note separates an observation from a diagnosis: “Clicks to this section declined; we are checking lost queries, demand changes and recent page edits” is more honest than immediately blaming a Google update.

Include material limitations: incomplete enquiry attribution, a changed consent setup, a broken event, a site outage or insufficient data. Flag a tracking discontinuity on the chart rather than silently treating the numbers before and after it as comparable.

A practical SEO report template for clients

Objective and scope: the service, audience and business outcome the work supports.
Executive summary: the main result, main problem and decision needed.
Performance comparison: dates, sources, relevant filters, clicks, impressions and useful landing pages.
Business outcomes: verified submissions, qualified opportunities and confirmed sales where available, with definitions.
Work completed: changed URLs, deliverables, implementation dates and verification.
Risks and limitations: measurement gaps, declines, unresolved issues and alternative explanations.
Next priorities: a short task list with an owner, intended outcome and dependencies.
Client decisions: approvals, subject expertise or development time required.

A dashboard can accompany this, but it should not replace the explanation. Agree a reporting rhythm that fits the work and sales cycle. Daily ranking movement rarely deserves the same attention as a broken enquiry journey.

Questions to ask at your next agency review

Which improvements came from non-branded searches relevant to our offer?
Which landing pages produced suitable enquiries, and how was that verified?
What was implemented this period, and what is still waiting?
Which numbers changed because of tracking or attribution settings?
What got worse, and what evidence supports the proposed response?
If we can fund only one next task, why is that the priority?

You should also know who owns the Search Console and analytics properties. Give the agency appropriate access rather than surrendering control of business accounts. A report should be reproducible from the agreed sources, not depend on an unexplained screenshot you cannot check.

Get a second opinion on your SEO reporting

If your report is hard to connect with business decisions, ask Craig for SEO advice. Share your website, the objective and a redacted report you are authorised to share. Remove customer details and confidential figures that are not needed. The conversation can focus on priorities, measurement gaps and what useful agency accountability should look like, without promising rankings or revenue.

Find related reading in the consultancy blog, or watch the SEO video answers for practical explanations from Craig's business and SEO channel.