SEO or GEO: where should the budget go in 2026?

Every quarter someone publishes a statistic showing search traffic collapsing, and every quarter someone else publishes one showing organic still driving most qualified leads. Both are measuring something real. This is the actual data as of mid-2026, what it does and does not justify changing, and how we would split a search budget on the strength of it.

What the measured data says

The Pew figures come from a panel study of 68,879 Google searches by 900 US adults in March 2025, of which 12,593 produced an AI summary. Percentages are of all visits, not of impressions — so they measure behaviour, not ranking. Nothing here is a European-market measurement; treat all of it as directional outside the US.

What genuinely changed

The click-through rate on informational queries has fallen hard, and the decline maps closely onto the AI Overview rollout rather than to any algorithm update. If your traffic model depended on ranking for "what is X" and monetising the visit later, that model is structurally worse than it was two years ago and will not recover.

What has not changed is the bottom of the funnel. Somebody comparing three vendors, checking a price, or looking for a supplier in their city still clicks, because the decision cannot be made inside a summary. The traffic that disappeared was mostly the traffic that was hardest to convert.

Why this is not an argument for dropping SEO

AI answers are assembled from pages that already rank. There is no separate index of AI-friendly content — being retrievable and being ranked are largely the same technical problem, which means the work that earns citations is the work that earns rankings. Abandoning SEO to fund GEO removes the input the AI layer is reading from.

The second reason is measurement. Classic search has fifteen years of tooling: Search Console, rank tracking, attributable clicks. GEO has manual prompt testing and a referral line in analytics. Shifting a majority of budget from a channel you can measure to one you cannot is not a strategy, it is a bet you will be unable to grade.

What GEO actually adds on top

GEO is worth funding for what it does differently: making your facts unambiguous and consistent so a model attributes a claim to you rather than to a competitor. Structured data, an llms.txt file, entity consistency across the web, and content shaped as clean extractable answers. Our GEO Starter is €599 one-time and does exactly that scope — the mechanics are covered in our GEO guide.

The realistic return is not traffic, it is presence. Being the business named when someone asks a chatbot for a recommendation is a brand outcome that shows up later as branded search and direct enquiries, not as a referral click you can attribute this week.

How we would split it

Foundations first, in this order: technical SEO and page speed, then the commercial pages, then GEO as an addition. Concretely that is SEO Starter at €499 before GEO Starter at €599, not instead of it. On a site where crawlability or Core Web Vitals are broken, GEO work is decoration on a foundation that cannot hold it.

If you already have solid technical SEO and rank reasonably, invert the priority: your marginal euro does more on structure, schema and answer-shaped content than on chasing another two positions for a keyword whose click-through rate is falling anyway.

Change the metric, not just the tactics

The most useful response to a falling click-through rate is to stop reporting on sessions. Track enquiries per month, branded search volume, direct traffic and referrals from AI platforms. Those four together tell you whether visibility is converting; impressions and average position increasingly do not.

It also changes what content is worth writing. A page that answers one specific question completely earns citations and converts the people who do click. A page written to hit a word count on a broad keyword now loses on both counts — see do websites still matter in the AI era for the wider picture.

Sources

Every figure in the table above comes from one of these studies, checked July 2026.

• Pew Research Center — Google users are less likely to click on links when an AI summary appears ↗

  Published 22 July 2025. 68,879 searches by 900 US adults; 8% click rate on a traditional result with an AI summary vs. 15% without, and 1% on links inside the summary.

• Search Engine Land — Google zero-click searches reach 68% in early 2026 (SparkToro study) ↗

  SparkToro clickstream analysis: 68% of US Google searches end without a click; the share of searches generating at least one click fell 9.51 percentage points between 2024 and 2026.