The Great Declick

For two decades, the deal in organic search was simple: earn a top ranking and Google sends you the click. Across a portfolio of business-to-business websites, that deal has quietly broken down.

We compared query-level Google Search Console data for the same calendar quarter one year apart (April–June 2025 versus April–June 2026) across 18 B2B accounts. Holding search ranking position constant, we found that click-through rates on non-branded queries have fallen sharply, while branded search is essentially untouched. The pattern is consistent with the rise of answer-style search results, AI Overviews chief among them, that resolve a searcher's question on the results page itself before any click occurs.

Ranking on page one no longer guarantees traffic for non-branded search. The queries most exposed are exactly those an answer engine can satisfy directly: definitions, how-tos, comparisons, and category research. Branded demand, where the searcher is looking for a specific company, remains a reliable source of clicks. For B2B marketing teams, that shift has direct consequences for content strategy, channel mix, and how organic performance should be measured.

Since answer-style search features and AI Overviews expanded across Google's results pages, marketers have traded anecdotes about “disappearing” organic clicks. Most public commentary on the topic relies on third-party keyword tools that model click-through rates rather than measure them, or on single-site case studies that can't separate a platform-wide shift from one site's idiosyncrasies.

We wanted to answer a narrower, more answerable question with first-party data: across a set of real B2B websites, did non-branded organic click-through rates actually decline year over year, and if so, for which kinds of queries?

Google Search Console reports impressions, clicks, click-through rate, and average ranking position for every query a site appears for. That makes it possible to compare the same query in two time periods and, crucially, to control for ranking position, so we are measuring a change in click behavior rather than a change in where a site ranks.

This is not a controlled experiment, and it cannot prove that any single SERP feature caused the decline. However, it can measure the size and shape of the change across many accounts, test whether the pattern holds broadly rather than in a few outliers, and identify the query types most affected.

Sample

The study covers 18 B2B accounts drawn from our agency's client base. All individual account names, domains, and identifying query details have been withheld; every figure in this report is reported in aggregate. In total the comparison set contains 5,159 queries that appeared in both time periods, of which 1,672 held a stable ranking position across the two periods and form the core of the analysis.

Time Periods

Google Search Console retains only the trailing 16 months of query data. As of mid-2026 that makes a clean pre-AI-Overview baseline (early 2024) unavailable. We therefore compared the second quarter of 2025 (April 1–June 30, 2025) against the second quarter of 2026 (April 1–June 30, 2026). Using the identical calendar quarter one year apart controls for seasonality, as B2B search patterns differ meaningfully between quarters, while still capturing a full year of change in how Google presents results.

One consequence of this window deserves emphasis: both periods sit after AI Overviews began rolling out broadly in 2024. This study therefore measures the continued expansion of answer-style search over a recent year, not a clean before-and-after.

Query Classification

Every query was tagged by intent using pattern rules: branded (contains a company's name), informational (definitions, how-tos, comparisons, guides, “best,” “vs,” and similar), navigational (logins, portals, contact and careers pages), or commercial. Queries with no informational or navigational signal were treated as commercial, because for these B2B sites the residual is overwhelmingly product and category terminology. Intent segmentation serves as the primary lens because informational queries trigger answer-style results at far higher rates than commercial or branded ones.

Controls

Three controls keep the comparison honest. First, position stability: we compare click-through rate only for queries whose average ranking position moved by no more than two places between periods, so a CTR change reflects click behavior rather than a ranking change. Second, a matched-query set: only queries present in both periods enter the year-over-year comparison, excluding terms that simply emerged or disappeared. Third, an impressions floor: a query must have at least 50 impressions in both periods to count, so no conclusion rests on thin data.

We deliberately did not filter accounts by click volume. Clicks are the very quantity under study; excluding low-click accounts would bias the sample toward high-CTR sites and understate the effect. Instead, all 18 accounts are included, small accounts simply contribute fewer qualifying queries, and findings are checked against a per-account view so that no single large account can drive a headline.

Non-branded click-through rates are collapsing

With ranking position held constant, average click-through rate on informational queries fell 38% year over year, from 0.90% to 0.56%. Commercial queries fell 20%, from 1.63% to 1.30%. Branded queries held flat, slipping 2% from 10.1% to 9.9%.

Non-branded queries where these sites rank on page one now convert impressions to clicks at roughly 1%. In other words, for the kinds of searches that make up the top of the B2B funnel — someone researching a problem, a category, or an approach — a top organic ranking now produces almost no traffic. The gap between branded and non-branded search has widened into a chasm: branded search converts at ten to twenty times the rate of everything else.

The zero-click surge

If click-through rate is the average, zero-click prevalence is the distribution. Among page-one queries with at least 100 impressions, the share earning near-zero clicks (under 1% CTR despite a first-page ranking) rose from 53% to 82% for informational queries in a single year. Commercial queries rose from 55% to 66%, and even branded queries edged up from 14% to 18%.

This is the single most consequential number in the study. By the second quarter of 2026, more than four in five informational queries where these B2B sites ranked on the first page of Google delivered essentially no traffic. The ranking still exists; the visit does not. For content built to capture informational demand — the blog posts, guides, and explainers that anchor most B2B SEO programs — page-one visibility has largely decoupled from page visits.

Seen more, clicked less: the AI tax

The mechanism becomes visible when impressions and clicks are compared directly. For informational queries, impressions actually rose 7% year over year while clicks fell 39%. Google is showing these pages as often as ever, but sending far fewer visitors. Commercial and branded queries show the same directional gap, with clicks falling faster than impressions in every intent category.

This divergence is the clearest signature of answer-style search in the data. When a results page resolves the query itself, through an AI Overview, a featured snippet, or a similar feature that often draws on the ranking page as a source, the impression is still logged, and may even increase as the page is cited, but the click never happens. The widening gap between impressions and clicks is, in effect, a tax levied on organic content: the content does the work of informing the searcher, and the platform keeps the visit.

How broad is the effect?

An average can hide as much as it reveals, so we tested whether these declines are broad-based or the artifact of a few large accounts. They are broad-based.

The informational CTR decline appeared in every one of the accounts with enough informational query volume to measure: a unanimous direction of the change. The commercial decline was present in 12 of the 18 accounts, a clear majority, though not universal: about a third of accounts held steady or improved on commercial terms.

Two further checks confirm the result is not an artifact of weighting or of one dominant account. Measuring each account's own click-through rate and averaging across accounts — so every account counts equally regardless of size — produces declines of the same magnitude (informational −41%, commercial −17%). And removing the single largest account from the calculation barely moves the headline figures (informational −37%, commercial −18%). The effect does not depend on how the data is weighted, and it does not depend on any one account.

The honest reading is that the informational decline is strong and consistent, while the commercial decline is a real majority trend rather than a universal law.

If ranking no longer reliably produces traffic on non-branded search, several long-standing assumptions need revisiting.

• •Both periods are post-AI-Overview. Because Search Console retains only 16 months of data, the 2025 baseline already reflects answer-style search. This measures a recent year of expansion, not a clean before-and-after, and the true multi-year impact is almost certainly larger than the figures here.
• •Correlation, not proven causation. We attribute the decline to intent-correlated zero-click and SERP behavior, with AI Overviews as the leading explanation. However, website content updates, commercial landscape volatility, and other factors could influence the performance as well.
• •Intent is rule-based. Classification uses pattern matching with a commercial default for unlabeled non-branded terms. This is accurate in aggregate but will misclassify individual edge cases.

Click-through rate by intent (position-stable)

Near-zero-click share of page-one queries

Year-over-year change: impressions vs clicks

Method summary: 18 B2B accounts · 5,159 matched year-over-year queries (1,672 position-stable) · same-calendar-quarter comparison, Q2 2025 vs Q2 2026 · position drift ≤ 2, minimum 50 impressions in both periods · first-party Google Search Console data.