Traffic down, revenue up: making the 2027 case for SEO in regulated markets

It's budget season, and one slide keeps turning up in planning decks: organic sessions, year on year, with a red arrow pointing down. If that slide is what your SEO budget gets judged on, you'll lose the argument for the wrong reasons.

The click loss itself is real. Ahrefs re-ran its AI Overviews study on December 2025 data and found that when an Overview appears, the top-ranking page gets 58% fewer clicks than you'd otherwise expect (Ahrefs). Their June 2026 numbers put the US figure at 68.8% (Ahrefs). And it isn't only position one. Position two lost roughly half its clicks, and position ten nearly a fifth (Business Wire). Ryan Law at Ahrefs is blunt about it: Overviews are "taking the majority of clicks that used to go to top-ranking pages."

We've been here before

Twice, by my count.

The first time was featured snippets. Google started putting an answer box above the results around 2014, and by 2017 Ahrefs had enough data to measure the effect. Across 2 million snippets, the first result's share of clicks fell from 26% to 19.6% when a snippet appeared, and the snippet itself took 8.6% (Search Engine Land). It wasn't all loss, either. Later click-through research found that positions two and three did better than average when the top result held a snippet (Search Engine Journal). Brands that won the box kept their visibility, even if they gave up part of the click.

The second was the slow creep of zero-click search. SparkToro put it at 50.33% of Google searches in 2019 and 64.82% across 2020 (SparkToro). Its 2024 study found 58.5% of US searches ended without a click, and only 360 of every 1,000 sent anyone to the open web (Search Engine Land). The three studies used different datasets, so I wouldn't draw a neat line through them. The direction is clear enough.

AI Overviews work the same way the snippet did, by answering on the results page. They just do it at a much bigger scale. The cut to position one is now close to three times what snippets took.

What NerdWallet shows

The best public example I know of a company going through this is NerdWallet, because it reports both numbers every quarter.

In Q4 2024, NerdWallet's revenue grew 37% while monthly unique users fell 20%. Management said the lost traffic was mostly non-monetising "learn" content, and that the product marketplaces were largely untouched (NerdWallet via Business Wire). The less flattering part came in 2025, when it raised performance marketing spend by 40%, to $417 million, partly to make up for lost organic reach (MarketBeat).

I take three things from that.

1. Informational clicks go first, and commercial ones hold up much better. Only around 10% of commercial or transactional keywords trigger an AI Overview (Big Leap).
2. You can replace organic reach with paid, but it costs real money. Budget cut from SEO in one planning cycle has a habit of coming back as media spend in the next.
3. Change how you measure before the budget conversation. If you walk in with sessions, the traffic chart makes the decision for you.

What it means in the four markets we work in

• iGaming and lottery: "how to play" and rules content is exactly what gets answered on the results page. Brand, bonus and registration searches still click through. I'd report first-time depositors from organic and branded search, and stop leading with sessions.
• Investing and trading: education hubs lose clicks, but they can win citations, and a regulated broker is the kind of source AI answers like to rely on. Track account openings and branded demand next to traffic.
• Health and treatments: healthcare has the highest share of keywords that trigger Overviews (Big Leap again), so expect the steepest informational losses here. Protect consultation and treatment pages first.
• B2B tech and SaaS: up to 72% of B2B buyers say they see AI Overviews in search (Search Engine Land). A lot of shortlists now get made without a click, so pipeline attribution matters more than sessions.

The slide I'd use instead

Gordon has written up how we measure the second row in How to measure AI visibility without fooling yourself. If your rankings dropped as well as your clicks, read his analysis of the 2026 core updates first, because that's a different problem. And if you'd like help building the 2027 case, talk to our AI visibility and SEO for YMYL teams.