SEO KPIs That Drive Decisions and What to Drop

Three audiences, three KPI sets — and the split most reports skip.

Why one KPI list always fails

Because three different people want three different things from it. Our guide to building an SEO report owns the document — its cadence, structure and audience. This page owns only which metrics go in it.

Your SEO executive needs to know whether last month's work is landing. Your marketing manager needs to know whether organic is contributing. Your CFO needs to know whether it is worth the retainer.

Put all of that on one slide and every reader skims past the two-thirds that is not theirs. The fix is not fewer metrics. It is three sets on three different cadences.

One scope note: all-channel marketing KPIs cover paid, social, email and organic together; everything below is organic search only.

The three KPI sets

Sort every candidate metric by the question it answers and the person who acts on it. If no one can act on it at that cadence, it is decoration.

Weekly — the SEO team (leading indicators)

• Indexation: pages submitted versus pages indexed, and the reason given for anything excluded.
• Non-brand impressions: whether more of your pages are becoming eligible for more queries.
• Queries entering positions 4–20: the band where a push is realistic. This is a working band we use, not a published benchmark.
• Crawl and coverage errors: new 404s, redirect chains, server errors, soft 404s.

None of these are outcomes. All of them move before outcomes do, which is exactly why the team watches them and the client does not.

Monthly — the marketing manager (outcomes)

• Non-brand organic clicks: demand you earned, separated from demand you already had.
• Conversions from organic: with the conversion definition written on the same slide.
• Page-level movement: which specific URLs gained or lost, not a site total.
• New pages earning their first impressions: the shortest honest proof that published work is working.

Quarterly — leadership and finance

One number: pipeline or revenue attributable to organic search, over a quarter, with the caveat attached rather than hidden.

The caveat is this. Organic rarely closes a deal on its own — someone finds you in search, leaves, returns via a brand search or an ad, and converts. Whatever your model credits organic with is a modelled figure, not a measured one. Say that out loud every quarter. Our explainer on how marketing attribution models work is the honest version of why the number moves when you change the model.

Brand vs non-brand is the split that matters

If you report one split and no others, report this one. Brand searches are people who already know you. Non-brand searches are people who do not. SEO's job is the second group.

A rising total-clicks line can be entirely brand, driven by a TV spot, a funding announcement or a paid campaign, while non-brand performance sits flat. The report looks like a win. Nothing SEO did caused it.

So non-brand impressions become the honest leading indicator: they rise when more of your pages become eligible for more queries that have nothing to do with your name.

Two things to know before you build the split in Search Console. Google's documentation on the Performance report's dimensions states that "Some queries are omitted from the report to protect user privacy" and that "They're included in chart totals, unless a query filter is applied" (Search Console Help — Dimensions and data groupings; the page shows no last-updated date, read 9 October 2026). Your filtered brand and non-brand figures will therefore not add up to your unfiltered total, and that gap is not an error.

Second, a brand filter is only as good as its spelling list. Include misspellings, the Hindi or regional transliteration, your product names and your founder's name, or they land in non-brand and flatter the number. The Google Search Console guide walks through the report itself.

KPI, audience, meaning, and how it gets gamed

Four metrics to stop reporting

Not because they are meaningless, but because as headline numbers they mislead the person reading them.

1. Average position as a headline. Google's own help page says the value is "the topmost position occupied by a link", averaged across all queries, and warns that it "is a complex metric that can be misleading if you don't understand the subtleties" (Search Console Help — impressions, position and clicks; no last-updated date shown, read 9 October 2026). Worth noting alongside this: Search Console's Generative AI performance report, which covers AI Overviews and AI Mode impressions, reports impressions only — no clicks, no queries and no position — and Google says of the surfaces it includes, "We expect to update this list over time" (Search Console Help — Generative AI performance report; rollout to all websites worldwide stated as of 31 August 2026, no last-updated date shown, read 9 October 2026). A headline position number already describes a shrinking share of how people see you.
2. Rankings for a handpicked keyword list. Whoever picks the list controls the result. Rank tracking is still useful diagnostically — see how rank tracking works — but it belongs in the SEO team's working view, not on the client's headline slide.
3. Raw organic traffic with no brand split. An organic traffic KPI without the brand cut cannot tell you whether SEO worked or whether marketing got louder.
4. Bounce rate as a quality proxy. A single-page visit where the reader got the answer and left is a success. Measuring it as failure pushes you towards padding pages so people have to click again.

The lag problem, and how to report through it

Organic work takes roughly three to six months to show meaningful movement. Google's SEO Starter Guide puts it more loosely: "Some changes might take effect in a few hours, others could take several months" (Google Search Central, last updated 10 December 2025).

Which creates a reporting trap. A monthly report that shows only outcomes will show nothing for the first two or three months of a new engagement, and nothing looks identical to failure.

Report leading indicators next to outcomes, and label which is which in the report itself. Two columns, headed leading and outcome, and a one-line note saying the first predicts the second with a lag. That single labelling choice is what stops a healthy month-three report from reading like a flat one.

What we put at the top of an SEO report

When there is no conversion tracking

Then you cannot report SEO's commercial value, and you should say so rather than substitute something that looks like it.

Traffic growth is not value. Neither is a rankings improvement, a position average or an engagement rate. Without a tracked form fill, call, enquiry or purchase, there is no honest line from organic search to money, and promising one in month one is how reporting credibility gets spent before it is earned.

So fix the tracking first. Get conversion tracking set up properly, agree on what counts as a conversion in writing, and until that is live report only what is true: indexation, non-brand impressions, non-brand clicks and page-level movement. State in the report that commercial reporting begins once tracking is live.

Where to take this next

Pick your three sets this week. Four weekly leading indicators for the team, four monthly outcomes for the manager, one quarterly number with its caveat for leadership.

Then delete something. A report that drops average position, the handpicked rank table and bounce rate is shorter, harder to argue with, and far more likely to get read to the end.