What Google's AI Overviews are actually doing to your organic traffic

A client called us in March worried her organic traffic had "fallen off a cliff" without any Google penalty, any redesign, or any drop in rankings. Position 3, same keywords, same content — just fewer people clicking through. She wasn't imagining it. The culprit sitting between her listing and the searcher was Google's AI Overview, and the data on what it's actually doing to clicks is now solid enough to plan around.

The number that made this real

The clearest evidence comes from a randomized field study run by researchers at the Indian School of Business and Carnegie Mellon University between January and February 2026. They recruited 1,065 Chrome users, split them into groups (normal search, AI Overviews hidden, and Google's AI Mode), and measured what people actually clicked. On the 42% of queries where an AI Overview appeared, outbound organic clicks fell from 0.61 to 0.38 per search — a 38% drop — and the share of searches ending with zero clicks at all rose from 54% to 72%. Just as notable: user satisfaction scores barely moved between groups, which the researchers took as evidence that AI Overviews are diverting traffic without actually serving searchers better.

It's not the flat cliff it looked like in 2025

The trend line isn't only going one direction, though. Seer Interactive's ongoing CTR tracking, updated in its "2026 Update" report, shows organic click-through rate on AI-Overview-affected queries bottoming out at 1.3% in December 2025 and climbing to 2.4% by February 2026 — an 85% recovery in two months, even as AI Overviews kept showing up on more searches. Nobody tracking this data thinks it's going back to pre-2024 levels, but it's a reminder that the current numbers on any given site are a snapshot of a system still settling, not a fixed new floor.

The one lever that still moves the number

The part of Seer's data most worth acting on is the citation gap. Pages that get cited inside an AI Overview receive roughly 20,700 clicks per million impressions, versus about 9,400 clicks per million impressions for pages that rank but aren't cited — more than double. Cited pages still trail queries with no AI Overview at all (around 33,500 clicks per million impressions), but the gap between "cited" and "not cited" is now bigger than the gap between "ranked #1" and "ranked #3" used to be. Getting summarized, not just getting ranked, is the new fight.

What we've changed in client work because of this

None of this changes strategy fundamentals — it changes where the effort goes first. On every content project since this data came out, we've moved these up the priority list:

• Answer the question in the first two sentences of a section. AI summarizers lift the most extractable sentence, not the best-written paragraph three lines down.
• Structure pages around one question per heading. A page with six clear question-shaped H2s gets cited more often than one long essay covering the same ground.
• Add FAQ and how-to schema markup where it's genuinely accurate — not stuffed in, but wherever the page actually answers a discrete question.
• Track citations, not just rankings. Rank-tracking tools now flag AI Overview appearances; if a page ranks well but never gets cited, that's the specific thing to fix.

We built this into the measurement system we set up during the first 90 days of a growth marketing engagement — citation tracking sits next to keyword rank and conversion rate now, not as a replacement for either.

If your traffic has quietly dipped without a ranking drop to explain it, that's usually the tell. Our growth marketing service starts by pulling this exact data for your own queries before we recommend a single change.