Rankings Are Not Revenue: Measuring SEO That Actually Matters

Keyword rankings are decoupled from revenue, and AI answers made it worse. Here is the measurement framework I use instead.

By Dalin de Graff

I stopped headlining keyword rankings in client reports. Not because rankings do not matter, but because the industry's favorite KPI has quietly decoupled from the thing clients actually pay for: revenue.

For years, "we moved you from position 8 to position 3" was the whole story. In 2026, that story can be fiction. You can climb three positions, watch traffic stay flat or fall, and have no idea why, unless you change what you measure.

The click funnel collapsed

The numbers are stark. SparkToro's June 2026 study, built on Similarweb's US desktop and mobile clickstream panel, found that 68.01% of Google searches in the first four months of 2026 ended without a click. In 2024 that figure was 60.45%. SparkToro calls it the fastest two-year acceleration of zero-click behavior they have ever tracked, and they attribute it largely to AI Overviews, which now appear on more than 20% of searches and reduce clickthrough rates by nearly 60%, per Ahrefs data cited in the study.

Pew Research independently confirmed the mechanism. In a study of 900 US adults' browsing behavior, Google users who encountered an AI summary clicked a traditional result in 8% of visits, versus 15% without one. They clicked the sources cited inside the summary in just 1% of visits. And they were more likely to end their browsing session entirely after seeing an AI summary.

Read that carefully. Moving from position 8 to position 3 means very little if both positions sit below an AI Overview that absorbed the query intent. Rankings still exist. They just no longer convert to traffic the way the old models assumed.

Rankings were always a proxy

Here is the uncomfortable part: rankings were never revenue. They were a proxy for revenue back when the proxy worked. A page ranking sixth for a high-intent commercial term routinely outperforms a page ranking first for a broad informational term, because intent beats position. As Awisee's 2026 measurement guide puts it, the strongest predictor of whether a page is working is revenue per session, not its rank.

AI search did not create this problem. It exposed it. The agencies still selling position reports as success metrics are selling a map of a landscape that no longer exists. The real visibility now happens above the rankings, inside AI-generated answers, and a rank tracker cannot see it.

Measure this instead: two layers

I now report on two layers, and rankings are a supporting detail inside the first one, never the headline.

Layer 1: business impact. This is the layer that survives zero-click. Organic revenue, qualified leads, calls, and conversion rate by landing page, pulled from Google Search Console and GA4. Pipeline contribution, not just last-click conversions. Wicked Reports makes the case well: measure revenue across the full journey with multi-touch attribution over 90 to 120 day windows, track assisted and first-touch conversions, and add a how-did-you-hear-about-us field to catch the influence tracking cannot see. SEO is assist-heavy and delayed. Last-click attribution credits whatever closed the sale and erases the content that started the journey weeks earlier.

Layer 2: AI visibility. This is the layer most teams are not tracking yet. How often is your brand named or cited across ChatGPT, Perplexity, Gemini, and Google's AI Overviews? Are competitors being recommended in your place? Track citation share on the commercial prompts your buyers actually ask, and watch branded search lift as a proxy for zero-click influence.

One caution on layer 2: a citation is not revenue. The correlation data on what predicts AI Overview visibility is instructive, Ahrefs analyzed 75,000 brands and found branded web mentions had the strongest correlation at 0.664, well ahead of domain rating or referring domains, as summarized by Awisee. That tells you brand building drives AI visibility. But a team obsessing over citation counts with no revenue tie-back is doing brand awareness, not SEO. Both layers need each other.

How reporting should change

Practically, this means rebuilding the monthly report. Lead with business outcomes: revenue, leads, calls, pipeline influenced. Follow with AI visibility: citation share, share of voice against competitors, branded search trends. Put rankings and impressions in the appendix where they belong, as diagnostic signals, not verdicts.

It also means having an honest conversation with stakeholders. When traffic falls but revenue rises, which SparkToro's Rand Fishkin notes is increasingly common, a rankings-first report reads as failure while the business is succeeding. When traffic rises but revenue is flat, the reverse. The report should explain the business, not flatter the agency.

I built First Rank's GEO department around this exact shift. On a recent GEO campaign we grew LLM search traffic 98% in 30 days, and the number that mattered in the report was not a ranking position. It was the business outcome attached to that visibility. That is the standard I hold every engagement to now: if a metric cannot be connected to revenue, leads, or demonstrable influence on a buying decision, it does not headline the report.

Stop asking "where do we rank?" Start asking "how much revenue did search generate, and where else did our brand show up?" Everything else is commentary.

Dalin de Graff

About the author

Dalin de Graff is the LLM + Search Engine Marketing Specialist at First Rank, where he built the agency's Generative Engine Optimization (GEO) department.

About Dalin de Graff ยท Contact