Activity, output and outcome are not the same thing

Confusion in SEO reporting almost always comes from mixing three layers of metric that measure different things. Activity metrics describe what the agency did: audits completed, pages written, links pitched. Output metrics describe what changed on the site or in search visibility as a result: rankings, impressions, indexed pages, Core Web Vitals scores. Outcome metrics describe what the business got: leads, bookings, revenue, retained customers. A report that only shows activity is a task list, not evidence. A report that shows output without connecting it to outcome tells you the machine is running, not that it is worth the fuel.

Good reporting shows all three, but weights outcome most heavily and is explicit about the lag between them — activity this month typically shows up as output next quarter and outcome the quarter after that, particularly for competitive commercial terms.

A measurement model from impressions to revenue

The funnel below is the shape we build for most clients, adapted to whatever the actual conversion path is — ecommerce checkout, lead form, phone call, or booked appointment.

Every stage of that funnel should have its own conversion rate tracked over time. If leads are flat but closed revenue drops, the problem is sales handling, not SEO. If clicks are flat but leads drop, the problem is likely on-page or landing page conversion. Reporting that only shows the top and bottom of this funnel hides exactly where an actual investigation needs to start.

Search Console vs analytics: expect discrepancies

Search Console and Google Analytics (or any other analytics platform) will never match exactly, and a discrepancy is not automatically a tracking error. Search Console counts clicks at the point a user leaves the search results; analytics counts sessions after the page loads and any tracking script fires, which means bounced page loads, ad blockers, and consent banner delays can all cause sessions to under-count relative to clicks. A gap in the 10–20% range is common and not inherently a red flag; a gap that suddenly doubles between two reporting periods usually is, and is worth investigating as a tagging or consent-mode issue before assuming a traffic problem.

Attribution limits for organic search

Last-click attribution systematically undercounts organic search’s contribution, because a large share of buyers research via organic search early in a decision and convert later through a branded search, a direct visit, or an email click. This does not mean organic gets a blank cheque of credit for everything; it means single-touch attribution models should be treated as a floor on organic’s contribution, not a ceiling, and multi-touch or assisted-conversion views should be checked before concluding that a channel “isn’t working” based on last-click numbers alone.

Branded vs non-branded split

Lumping branded and non-branded search together in a report flatters agencies and misleads owners. Branded search growth mostly reflects offline marketing, word of mouth and existing brand awareness — it is a poor way to judge whether an SEO programme is doing new-demand generation. Non-branded search growth is the more honest signal of whether the programme is winning new visibility among people who did not already know the business existed. Every dashboard should report these two separately, not as a combined “organic traffic” number.

Share of voice as a leading indicator

Share of voice — your estimated visibility across a defined basket of relevant keywords compared with named competitors — is one of the few metrics that moves faster than revenue and is harder to game than a single rank tracker screenshot, since it aggregates across many terms rather than resting on one. Tracking it monthly against two or three real competitors, not a vague “industry average”, gives an owner a directional read on competitive position well before it shows up in the bank account.

Leading indicators in the first 90 days

Page-group and cohort reporting

Reporting rankings or traffic for the whole site hides where growth is actually coming from. Grouping pages by type — service pages, location pages, blog content, product pages — and reporting each group’s trend separately shows which parts of the strategy are working and which are dead weight. Cohort reporting, tracking a specific batch of pages published in a given month over time, is the cleanest way to judge whether new content production is paying off, since it isolates that batch from the noise of the rest of the site.

Honest forecasting

Forecasting organic growth should always be presented as a range with stated assumptions, not a single confident number. A forecast that says “we expect 20–35% growth in non-branded organic sessions over two quarters, assuming no major algorithm update and consistent publishing cadence” is honest. A forecast that says “you will get 500 new leads a month by June” without any stated assumptions is a sales pitch dressed as a projection, and it is the single most common thing that erodes client trust when it does not land.

The report that actually protects an SEO budget is not the one with the most charts. It is the one that says plainly which numbers moved, which did not, and what the agency is doing differently next month because of it.

Other red flags in agency reporting

Beyond vanity metrics and withheld access, watch for reports that never mention a metric going down, since organic performance is never a straight line up and a report that hides every dip is curated rather than honest. Watch for reports that change the reporting period or comparison baseline whenever the numbers look unfavourable — comparing this month to a cherry-picked low month rather than the same month last year, for example. And watch for any report that presents correlation as proof: “we published 10 blog posts and traffic went up” says nothing about causation without checking what else changed in the same window, including seasonality, algorithm updates, and competitor activity.

Frequently asked questions

Reporting cadence and who reads what

One report cannot serve three audiences. The owner wants to know whether the investment is producing pipeline. The marketing manager wants to know which pages and clusters are moving so they can plan the next quarter. The person doing the work needs granular diagnostics that would bore everyone else. Trying to satisfy all three in a single PDF produces a document nobody reads carefully, which is how underperformance stays hidden for two quarters.

Split it. The owner gets a one-page monthly summary: non-branded impressions and clicks, qualified enquiries from organic, cost per enquiry compared to paid, the two or three things that shipped, and the two or three things shipping next. No jargon, no screenshots of tool dashboards, no vanity charts. The manager gets a working review every month with page-group performance, cluster-level movement, content and link deliverables against plan, and any technical issues found. The practitioner keeps a live diagnostic dashboard that nobody needs to see in a meeting.

Cadence matters as much as content. Weekly reporting on organic search is noise — the underlying data is too volatile and the work is too slow to show weekly deltas. Monthly is right for review, quarterly is right for strategy. Reserve out-of-cycle communication for genuine events: a core update, a ranking drop, a site migration, a tracking break. If an agency sends a lengthy automated report every Monday and nothing else, that is volume standing in for insight.

Finally, agree the definition of every metric in writing before the first report. What counts as a qualified enquiry, whether phone calls are tracked, which subdomains are in scope, and whether branded terms are excluded from the headline number. Most reporting disputes are not disagreements about performance; they are disagreements about definitions that were never settled at the start.