SEO reporting: the complete guide

Most SEO reports fail before the second page. Not because the work behind them was bad — because the document was built for the wrong reader, in the wrong order, with no opinion about what matters. The standard agency report is a dashboard export with a cover page: every chart the tool can produce, presented at equal volume, with the implicit instruction figure out for yourself whether this was a good month. Clients don't figure it out. They skim it on a phone between meetings, register that it looks thorough, and archive it.

Three failure modes account for nearly every skimmed report we've seen:

• Written for the sender. The report documents activity — links built, pages optimized, audits run — because activity is what the agency can see. The client paid for outcomes, and the outcomes are either missing or buried under the activity log.
• No hierarchy. Twenty metrics presented with equal weight means no metric matters. Deciding what goes first, what goes big, and what gets cut is the editor's job — and in a report, you are the editor.
• No narrative. A chart says what happened. Only a sentence can say why it happened and what you intend to do about it. Charts without sentences are homework, and clients do not do homework.

The stakes are higher than most agencies admit. Between calls, the report is the only proof of life the retainer produces — it is the product, as far as the client's memory is concerned. A client who stops understanding the report starts shopping, usually right around the time a competitor's cold email lands promising "transparent reporting."

Before a single metric is chosen, answer one question: who actually opens this? In our experience there are three readers of an SEO report, and they want almost nothing in common.

The owner-operator reads for money. They will give you three sentences and one number, and they want both to be about leads, calls, and revenue set against the invoice you send them. Every keyword chart between them and that answer is a withdrawal from your credibility account. The CMO or marketing director reads for ammunition — they have a boss too, and your report is raw material for their report. They want progress against the plan you agreed, context against competitors, and a story clean enough to forward without editing. The technical lead or in-house SEO is the only reader who genuinely wants the keyword table, the crawl issues, and the change log — and the only one who will catch you padding it.

The practical implication is not three reports — it's one report, layered. Open with the executive summary written for the most senior reader. Follow with the scoreboard, then the proof, then the plan. Put the tables in an appendix. Never make an owner scroll past a crawl-error chart to find out whether the month was good; never make the technical lead hunt for the data that backs your claims. Each reader should hit their layer and stop, satisfied.

If you're in-house rather than agency-side, nothing changes except the names: reporting up is the same exercise, and your CMO is the client. The discipline of leading with outcomes matters even more internally, because your report eventually reaches someone who approves budgets and has never heard of a featured snippet.

A metric earns a place in the report by changing a decision or defending a budget. By that standard, the list is shorter than most reports suggest:

• Organic conversions and their value. Form fills, tracked phone calls, purchases, attributed revenue. This is the metric the invoice is judged against — it goes first, always.
• Organic traffic to money pages. Sessions to the pages that convert — service pages, product pages, pricing — not the site-wide total, which blog fluctuations can swing in either direction without a dollar changing hands.
• Keyword movement with revenue context. Movement on terms mapped to converting pages. "Up 12 positions on a keyword that feeds the booking page" is a result; "up 12 positions" alone is trivia.
• Share of voice against named competitors. Clients think in rivalries, not indexes. Visibility versus the three competitors they actually lose deals to is a metric they will lean into the screen for.
• AI visibility. Citation rate and presence across the AI engines — covered in full in the next section.

Then there's the vanity layer — metrics that survive in reports because they're easy to export and usually go up:

• Rankings without revenue context. Position three for a term nobody converts on is decoration. Rankings belong in the report only when they are attached to a page and a business outcome.
• Raw impressions. Inflated by every algorithm shift, every irrelevant query, every SERP layout change. An impressions chart that climbs while leads stay flat teaches the client to distrust your charts.
• Domain Authority. A third party's estimate of an input, not an outcome — and not a Google metric, however often it's presented like one. Reporting DA movement as a result invites a conversation you will lose.
• Total keywords ranked. "You now rank for 4,000 keywords" is a database artifact. Most of them are accidental, irrelevant, or page-nine noise.

The working rule: every metric must survive the "so what?" test twice. Impressions are up. So what? More people see us. So what? If the second answer isn't a business outcome, the metric goes to the appendix or goes away. Agree on a scoreboard of four to six metrics at kickoff, report those against goals every month, and resist additions — a scoreboard that grows is a scoreboard losing its argument.

The most convincing unit in an SEO report is not the keyword. It's the page. "We rebuilt your emergency-service page in March; it now ranks in the top three for the queries that matter, traffic to it has grown since, and it produced this month's tracked calls" — that is a sentence a client repeats in their own management meeting. The chain runs keyword → page → conversion, and every win in the report should be expressed as a link in that chain.

The reason page-level proof lands is that clients don't know what a keyword is worth — but they know exactly what their services page is worth. Anchoring results to URLs the client recognizes converts your work from abstraction into property: their page, performing. It also makes the report auditable, which builds the kind of trust that survives a bad quarter.

The honest obstacle is plumbing. The chain's three links live in three tools: Search Console holds queries and pages, GA4 holds behavior and conversions, the rank tracker holds movement — three logins, three exports, and an evening of spreadsheet stitching per client. This is exactly the join Aergos was built around: page-level analytics with keyword-to-page drilldowns, pulling Search Console, GA4, and keyword data into one view, so the keyword → page → conversion chain is something you read rather than something you assemble.

One more link completes the chain for lead-gen and local clients: the conversions themselves. A dental practice's conversions are phone calls; a contractor's are form fills and calls. Aergos's first-party tracking SDK attributes phone and form conversions back to the organic pages that produced them — which means the report can say "this page produced these leads" instead of "this page got traffic, and we assume the rest."

Your clients are already typing their own brand into ChatGPT and noticing who comes up. "Do we show up in ChatGPT?" has joined "why aren't we number one on Google?" as the question every agency gets — and the agencies that put AI visibility in the monthly report before being asked own that conversation. The ones that wait get asked about it, which is a much weaker position.

Three numbers make AI visibility reportable rather than anecdotal:

• Citation rate. Across a tracked set of queries that matter to the client's business, the share where AI engines cite or mention the client. This is the AI-era analogue of "average position" — one number, trended monthly.
• Share of voice. When the client isn't the answer, who is? Presence versus named competitors inside AI answers turns an abstract channel into the rivalry framing clients already think in.
• Per-engine presence. ChatGPT, Perplexity, Gemini, and Google AI Overviews select sources differently and disagree with each other constantly. A blended average hides the story; the per-engine view shows where the client wins, where they're invisible, and where one engine's gap is next quarter's work.

For the non-technical reader, resist the urge to explain retrieval. One chart and one plain sentence: "When potential customers ask AI tools about [their service] in [their market], you're cited this often — here's the trend, and here's who appears when you don't." Frame it as a leading indicator: AI presence today shapes where buyers start their research tomorrow, and the work that earns it — being the clearest, most-cited source on the topics that matter — is the same work driving the rest of the report.

Mechanically, this is what Aergos AI visibility reports produce: citation rate, share of voice, and per-engine breakdowns, with narrative summaries written in plain language — built to drop into a client report rather than to require a translation layer.

The most valuable component of any SEO report costs nothing to produce and is missing from most of them: three sentences at the top, written by a human with an opinion. What changed. Why. What's next.

1. What changed — the single most important movement of the month, stated with a number. "Organic leads grew this month, led by the two service pages rebuilt in April."
2. Why — the diagnosis, in plain language. "The rebuilds reached the top three for their core queries, and call tracking shows those pages converting." If the why is partly unknown, say so — "we're still confirming the cause" reads as competence, not weakness.
3. What's next — the plan, with a date. "Next month the pricing page ships, and we expand AI visibility tracking to two new competitors."

The rules that keep the summary honest: write it last, after the data is in, so it's a conclusion rather than a hope. Write it for the most senior person who might ever see the report. Name at least one number. Use no jargon a client would have to look up — "crawl budget" and "SGE" have no business in an executive summary. And never blame mystery: "Google did something" is the sentence that teaches clients you don't know how your own channel works.

For most readers, the summary is the report — everything below it is the appendix to those three sentences. That's not a failure of the rest of the document; it's the point. The detail exists so the summary can be trusted, the way footnotes exist so a claim can be checked. Most months, nobody checks. The months somebody does, the depth is what saves you.

One report cannot serve three time horizons. The cadence that works runs on three layers, each with its own depth and delivery:

• The weekly pulse. Five lines, no meeting: notable movement, anomalies caught, work shipped. Its job is presence — proof between reports that someone is watching the account. A live dashboard the client can glance at does the same job passively; a branded portal does it best, because the glance happens under your logo.
• The monthly deep-dive. The report this guide is about: narrative, scoreboard, page-level proof, AI visibility, losses, plan. This is the retention artifact — the document that gets forwarded to the boss and re-read before the renewal call. It deserves real hours.
• The quarterly strategy review. Zoom out: progress against the original goals, market and competitor shifts, what the AI-search transition means for this client, next quarter's bets. Deliver it as a meeting with a deck, not an email — strategy discussed out loud renews retainers; strategy attached as a PDF gets skimmed like everything else.

On format: live portals beat PDFs for the pulse, because freshness is the entire value. For the monthly report, a document — PDF or deck — still wins, for an unglamorous reason: it gets forwarded. Your report will be attached to an internal email, dropped into a board pack, pulled up in a budget meeting where you aren't in the room. A login can't travel; a document can. The strongest setup is both: the portal for any-time access, the document for the monthly moment.

Calibrate to the retainer. A small local-SEO engagement does not need a weekly pulse and a forty-slide quarterly — over-reporting eats margin as surely as over-servicing. A large engagement under-reported is a churn risk wearing a friendly face. The cadence is part of the commercial design, not an afterthought.

For most clients, the monthly report is the only artifact the agency relationship physically produces. It should look like the agency they hired — not like the toolstack the agency rents. Every third-party logo on a client deliverable quietly raises a question you don't want raised: what exactly is the agency adding, and could we just buy the tool?

White-labeling has levels. The floor is a logo dropped on a PDF export. The real version is the full surface under your brand: your logo, your colors, and — the piece most agencies skip — your domain. When a client logs into reports.youragency.com and sees your brand on every chart, the reporting infrastructure reads as something your agency built. That perception is worth real money at renewal time, and it compounds: every weekly glance at the portal is a brand impression you own.

This is a first-class feature in Aergos rather than an add-on: white-label reports and a branded client portal — your logo, your color scheme, your custom domain — so every report and every login happens under the agency's name. And because Aergos pricing is flat, white-labeling doesn't carry the per-client surcharge that punishes you for growing; the economics of adding your fifteenth client look like the economics of your fifth.

White-labeling the reports is one piece of a larger play — reselling SEO under your own brand end to end. That full model, including packaging and pricing, has its own guide.

Fully automated reports have a smell, and clients pick it up fast: the summary that says "traffic increased 4.2% month over month" in the same tone every month, the insights that restate the chart beneath them, the recommendation that doesn't quite fit this client. Automation applied to the wrong layer doesn't save the relationship time — it spends the relationship down.

The split that works is clean:

• Automate the data layer completely. Pulling Search Console, GA4, rank, and AI visibility data; assembling charts; populating the scoreboard; flagging anomalies. There is zero human value in copy-paste, and every hour spent on it is an hour stolen from the three sentences that matter.
• Keep the narrative layer human. The executive summary, the diagnosis of why things moved, the judgment call on what next month's priority is. This is the part the client is actually paying a human to do — and the part where a template is instantly detectable.

Done right, automation doesn't replace the narrative — it funds it. When the data assembly takes minutes instead of an evening, the saved hours go into the thinking, and reporting night stops being the worst night of the agency month. Aergos is built for exactly this split: the data layer assembles itself, AI visibility reports arrive with draft narrative summaries, and the whole reporting workflow is MCP-driveable — your AI assistant can pull the report data and draft the narrative for a human to review, edit, and sign.

Bad months are when reporting earns its keep. The instinct — every agency feels it — is to bury the dip: lead with whatever went up, shrink the declining chart, hope the client doesn't ask. It fails on both branches. If the client notices, you've converted one bad month into a credibility problem that outlives it. If they don't notice, you've taught yourself that the report is a marketing document, and the next bad month gets buried deeper.

The play that builds trust is the opposite, run in three moves:

1. Lead with the loss. It goes in the executive summary, named plainly, with the number. The client should never discover a decline on page six that the summary didn't mention.
2. Diagnose it. Algorithm update, seasonality, a tracking change, lost links, a competitor's new page — say which, and show the evidence. If you don't know yet, say that: "we're investigating, and here's what we've ruled out" reads as control. The one unacceptable diagnosis is silence.
3. Ship the plan. What you're doing about it, who owns it, and when the client should expect to see the line move. A loss with a dated plan attached is a project; a loss without one is a crisis.

A diagnosed loss reads as command of the channel. A hidden loss, when discovered — and they are discovered — reads as either incompetence or deceit, and the client gets to pick. In our experience, the agencies that report losses plainly keep clients longer than the agencies that only ever report wins, because the client learns that when this agency says things are good, things are actually good. That's the asset a decade of green-arrow reports never builds.

The cheapest version of this play happens before any bad month arrives: pre-wire expectations. The report that says, in a good month, "organic is volatile; we expect dips and here's how we'll respond when one comes" makes month seven's dip a confirmation of your competence instead of a test of it.

Everything above compresses into one structure. This is the monthly report we'd defend — seven sections, in an order that is itself the message:

A few of the load-bearing details. The executive summary is written last and placed first — it's a conclusion, not an introduction. The scoreboard shows the same four to six metrics every month against the goals agreed at kickoff; consistency is what makes month-over-month meaningful, and the discipline of not adding metrics is what keeps it readable. Wins are expressed as keyword → page → conversion chains anchored to URLs the client recognizes — one well-evidenced win beats five vague ones.

The AI visibility snapshot stays small on purpose — one chart, one sentence — until the client's interest grows it. Losses get their own section rather than a euphemism inside "opportunities"; the section exists every month, even when it says "nothing material declined," because its permanence is what makes it credible. Work shipped comes after outcomes, deliberately: activity justifies itself by pointing at the results above it, never the other way around. And the appendix takes everything else — full keyword tables, technical detail, raw trends — where the technical reader can verify and nobody else has to scroll.

Outcomes before activity. Narrative before tables. Losses beside wins. Get the order right and the same data that used to get skimmed starts getting forwarded — which is the only metric of reporting success that matters.