Fewer clicks, higher prices: the new economics of Google search ads

Executive summary

Alphabet's Search & other advertising revenue reached $63.3 billion in the second quarter of 2026, up 17% year over year from roughly $54.2 billion a year earlier (Confidence: High — corroborated by 9to5Google and MarketScreener, both citing the official earnings release). That's not a company in decline, but sitting next to that number are two others that do not fit the same story: the average cost per click across US Google Ads campaigns hit $5.42 in the 2026 WordStream/LocaliQ benchmark report, and Search Engine Land's own reporting independently confirms that figure is "more than double what it was a decade ago" (Confidence: High — primary report, corroborated independently). Meanwhile eMarketer forecasts Google's US ad revenue will grow just 5.6% in 2026 — the slowest of the "big three" platforms — and projects Google's share of US search ad spend will fall to 48.5%, below 50% for the first time in more than twenty years (Confidence: High — confirmed via eMarketer and multiple secondary outlets).

Those two facts are not actually contradictory, and reconciling them is the point of this piece. Google is extracting more revenue per unit of advertiser attention even as the total pool of attention it commands shrinks relative to competitors — chiefly Amazon, whose US ad revenue is forecast to grow 17.9% in 2026 to $56.71 billion, more than three times Google's growth rate (Confidence: High). Advertisers aren't leaving Google; they're paying more per click, bidding into a Google-designed system (AI Max for Search) that has expanded the query space they're forced to compete for, while diverting incremental dollars to retail media and other channels. Digiday's advertiser-side reporting captures the mood bluntly: Adthena's chief marketing officer told the outlet, "The CPC pain is real" (Confidence: High — verbatim quote), while a Mediaplus media partner separately confirmed, "CPC inflation is definitely happening to a number of different advertisers" (Confidence: High).

The clearest corrective to the "Google is forcing everyone onto AI Max whether they like it or not" narrative: Google's own no-opt-out migration of remaining Dynamic Search Ads campaigns onto AI Max — originally slated for September 2026 — has been delayed to February 2027 after advertiser pushback, and Google reopened DSA campaign creation on June 15, 2026 (Confidence: High — Search Engine Land, corroborated by SEO Roundtable, PMG, and a Google blog post). That delay is more revealing than the original plan would have been — it shows advertiser resistance still caps how aggressively Google can push AI-native ad formats, even as its own economics keep improving.

A timeline / comparison table

Part 1 — The revenue number that doesn't match the mood on the ground

Alphabet's Q2 2026 print is, on its face, one of the strongest quarters the company's ad business has reported in years. Search & other advertising revenue came in at $63.3 billion, up 17% from about $54.2 billion in the year-ago quarter, with retail and finance the largest contributors to growth (Confidence: High). Total Alphabet revenue reached $119.8 billion, up 24% year over year, with operating income around $40.8 billion; YouTube ads revenue climbed to $11.06 billion from $9.79 billion (roughly 13% growth); and Google Cloud grew 82% to $24.77 billion (Confidence: High — cross-checked against CNBC and Yahoo Finance). Google also leaned hard on user-growth numbers in its own messaging: AI Mode surpassed 1 billion monthly active users, and the standalone Gemini app reached 950 million, up from roughly 750 million in February 2026 (Confidence: High — confirmed across four named outlets: Digital Information World, Unite.AI, Hypertext, AndroidHeadlines). Read in isolation, this looks like a company converting the generative-AI transition into ad revenue, not the disrupted incumbent AI search was supposed to produce.

The friction starts when that revenue is set against the eMarketer forecast published five months earlier, in February 2026, projecting only 5.6% growth in Google's US ad revenue for the full year — the slowest of Google, Amazon, and Meta — versus Amazon's projected 17.9% growth to $56.71 billion and Meta's 14.2% (Confidence: High). A single quarter's 17% growth sitting well above a full-year forecast of 5.6% isn't necessarily a contradiction — forecasts get revised, and one strong quarter doesn't overturn an annual trajectory — but it's a real gap between what analysts expected in February and what printed in July, and it should be treated as an open tension rather than quietly reconciled. Either the forecast undershot, or growth is front-loaded and due to decelerate later in the year; the source material available doesn't resolve which.

Part 2 — What CPC data actually says, and what got fabricated in early drafts of this story

A widely circulated claim that "Google CPCs fell to $2.96" does not exist in the actual data, and it's worth correcting on the record. The 2026 WordStream/LocaliQ benchmark report — 13,474 US campaigns, April 2025 through March 2026 — puts the cross-industry average Google/Microsoft Search CPC at $5.42, describing the year-over-year change as "fairly stable" or "mellow" rather than a large across-the-board move (Confidence: High — directly fetched primary report). Search Engine Land's independent analysis arrives at the same $5.42 figure through its own methodology and calls it "more than double what it was a decade ago" (Confidence: High — independent corroboration, satisfying two-source verification). Individual verticals diverge sharply beneath that average: Real Estate CPCs rose 27.27% year over year in the WordStream data, Education fell 22.79% — "average CPC" flattens genuinely different advertiser experiences.

Other trackers cited within Search Engine Land's reporting show steeper movement than the WordStream headline number: Stackmatix data put Google Search CPC growth at 14–18% year over year, with LinkedIn even higher at 18–22%; some accounts reportedly saw up to 25% inflation on primary commercial keywords, per Digiday's reporting cited within the same piece; and AdThena data showed advertiser participation in search auctions up 35% year over year (Confidence: High — secondary citations within the directly fetched article, not independently re-verified against primary vendors). Ben Wood of Search Engine Land put it this way: "Rising CPCs aren't just the result of more competition inside search auctions. They're increasingly the result of what happens before anyone places a bid" (Confidence: High — verbatim quote).

Digiday's advertiser-side reporting, published almost exactly one year after AI Max for Search launched, adds agency-level texture: Mediaplus clients saw average CPC increases of 10–15% year over year tied specifically to AI Max; Collective Measures clients saw typical increases around 10%, up to 25% for some advertisers, alongside 7–10% year-over-year search spend growth; Go Fish Digital clients saw 15% year-over-year search spend growth (Confidence: High — directly fetched, all figures and firm attributions confirmed). Lauren Beerling of Collective Measures offered a mechanism for why: "AI Max is also opening up query space our clients weren't previously bidding on" (Confidence: Medium — quote not independently verified against a second transcript). That's the crux of the CPC story: AI Max's broad-match-style expansion doesn't just raise prices on existing keywords, it pulls advertisers into auctions for queries they never targeted — mechanically inflating both spend and the auction-participation figures AdThena measured above.

Counterpoint worth taking seriously: not every account is seeing pain. Go Fish Digital's David Dweck told Digiday, "AI Max will be the most widely adopted ad type in probably one to three years" (Confidence: High — verbatim quote) — a bet that broader query coverage will pay off net-positive once advertisers adapt bidding strategy, even if the transition period is expensive. The disagreement here isn't about whether CPCs are rising — everyone agrees they are — it's about whether that's a cost of a genuinely more efficient auction or a toll imposed by Google's design choices with no efficiency gain attached.

Part 3 — The AI Overview click story is messier than "collapse," and that matters

Seer Interactive's November 2025 study is the single most-cited data point behind the "AI Overviews are killing paid clicks" narrative: paid CTR on AI-Overview-present queries fell 68%, from 19.70% in June 2024 to 6.34% in September 2025, based on 3,119 informational queries across 42 organizations (Confidence: Medium — headline figure; requires the follow-up correction below every time it's cited). Seer's own follow-up study, published roughly five months later by the same research team, found that paid CTR on AIO-present queries had recovered to 16.21% by February 2026, up from 14.64% in January 2025 — meaning the steep decline had already substantially reversed before this article was written (Confidence: Medium). Citing the 68% collapse without the follow-up is a claim that won't survive its first fact-check; the honest version is "sharp initial suppression, partial recovery," not "structural collapse."

AdThena's separate analysis of more than 5 million ads across six industries, run December 2025 through January 2026, adds industry texture rather than a national CPC trend: roughly one in four paid ads were found to be impacted by AI Overviews, with industry content mixes varying substantially — Telecom, Technology, and Retail queries dominated by comparison content; Healthcare queries 74% news-driven; Financial Services 54% FAQ-driven; and in Healthcare specifically, ads appeared below the AI Overview about 64.6% of the time (Confidence: Medium — top-line scope confirmed; two vertical-level CPC sub-claims from earlier drafts couldn't be verified and were dropped).

Note for the design team: a two-panel chart works well here — left panel: Seer's CTR line from 19.70% (June 2024) down to 6.34% (Sept 2025) and back up to 16.21% (Feb 2026), to visually correct the "collapse" narrative in one image. Right panel: AdThena's industry-by-industry AIO-impact share, to show the variance underneath any national average.

Part 4 — The forced migration that got walked back

Google Marketing Live 2026, held May 20–21, positioned AI Max and AI Mode as the future default of how Search ads work. Google's own figures from the keynote: AI Max for Search delivers 27% more conversions compared to manual campaigns; for every $1 spent on Google Search globally, Google claims an incremental $6 return at scale; AI Overviews reached 2.5 billion monthly active users; AI Mode queries average three times the length of traditional searches and have been doubling in volume every quarter since launch; and Google plus YouTube claim presence in 82% of consumer discovery journeys and 89% of purchase journeys (Confidence: High that these are Google's stated claims, from the official Think with Google recap — but they're self-reported marketing figures, not audited numbers). Chief Business Officer Philipp Schindler told the room, per Mi3's coverage, "We have made a decade's worth of innovation in the last year alone... AI gives search superpowers" (Confidence: Medium — named trade outlet, not verified against a primary transcript).

An independent counterweight belongs in the same paragraph, and it doesn't need to be invented: the agency-level CPC increases documented above (Mediaplus's 10–15%, Collective Measures' up to 25%) are the advertiser-side reality sitting underneath Google's platform-level "27% more conversions" framing — the same campaigns can be simultaneously converting better on Google's own attribution and costing meaningfully more to run, which is exactly what a broader, more auction-competitive query space would produce.

Set against that gap, Google's original plan to force all remaining Dynamic Search Ads campaigns onto AI Max with no opt-out, starting September 2026, would have looked like leverage without consent. It didn't happen on that timeline. Search Engine Land reported — corroborated by SEO Roundtable, PMG, and a Google blog post — that the full forced migration has been delayed to February 2027 following advertiser pushback; only automatically-created-assets configurations and campaign-level broad-match settings still transition on the original September 2026 schedule, and Google restored the ability to create new DSA campaigns on June 15, 2026 (Confidence: High). Whatever revenue and adoption numbers Google presents at its own keynotes, the delay shows that advertiser resistance still sets a real ceiling on how fast Google can push its highest-margin, AI-native ad format.

Part 5 — Google's shrinking share of a growing pie

The least intuitive fact in this entire dataset is also the best-corroborated one: even as Google's Search revenue grows, Google's share of the market it dominates is shrinking. eMarketer's February 2026 forecast puts Google's share of US search ad spend at 48.5% for the year — the first time below 50% in more than two decades — with Amazon capturing the largest share gains (Confidence: High). Total US search ad spend climbed 9.9% in 2025 to roughly $142 billion, and Amazon's share of that spend is projected to reach 25.7% by 2027; retail-media-search specifically grew 17.7% in 2025 to $38 billion (Confidence: High — all figures directly confirmed from the primary eMarketer article). WARC's global ad forecast shows the whole industry accelerating: 2025 growth was revised up to 7.4% (to $1.17 trillion), described as WARC's first positive revision in over a year, with an 8.1% forecast for 2026 in the same report — though later WARC updates push the 2026 figure as high as 9.1%, meaning the forecast keeps moving upward over time (Confidence: Medium — corrected from an earlier draft that mislabeled which year 7.4% applied to).

On the demand side, Digiday+ Research found that 37% of brand and agency professionals reported their companies had seen decreases in upper-funnel or informational search traffic attributable to AI, versus 32% reporting no change (Confidence: Medium — directly confirmed figure), and Forrester projects brands may cut open-web display spend with publishers by 30% in 2026 in favor of CTV, streaming audio, and paid social (Confidence: High — confirmed across four named trade outlets citing the same Forrester prediction). Put together: the same forces that are suppressing organic and informational traffic are pushing ad dollars toward retail media, CTV, and paid social — channels where Amazon in particular is gaining relative to Google.

What everyone is missing

The revenue and the market-share numbers are both true, and neither one is the whole story. Coverage of Google's ad business tends to pick one number — 17% Search revenue growth or sub-50% market share — and use it to declare either "Google is fine" or "Google is losing." Both readings miss that Google can extract more money per unit of advertiser attention while losing relative share of total ad spend to Amazon and retail media. Those are compatible facts, not competing ones; a strategy built on only one of them is built on half the picture.

CPC inflation is partly mechanical, not just competitive. The instinct is to read rising CPCs as more advertisers bidding for the same fixed inventory. AdThena's 35% year-over-year growth in auction participation, combined with Collective Measures' observation that AI Max is "opening up query space clients weren't previously bidding on," suggests a different mechanism: Google's own broad-match expansion is manufacturing more auctions for advertisers to lose, raising average CPC even when keyword-level competition hasn't changed much. That matters for budget planning — CPC inflation isn't a market condition advertisers can wait out, it's a product design choice Google controls the pace of.

The DSA delay is the most honest data point in this whole story, and it's underreported. Every other figure here is a number someone wants reported a certain way — Google favorably, advertisers sympathetically. The delay of the forced AI Max migration from September 2026 to February 2027, under direct advertiser pressure, is neither — it's Google backing off a plan because enough of its customer base pushed back hard enough. That's rarer and more diagnostic than any keynote statistic, and it should carry more analytical weight than it currently gets in trade coverage.

"AI Overviews are killing paid clicks" is already out of date. The 68% CTR collapse from Seer's November 2025 study is the number still circulating in most advertiser conversations, but Seer's own follow-up data shows more than half of that decline had reversed by February 2026. Using the collapse figure without the recovery figure isn't just incomplete, it actively misleads advertisers into overreacting to a phase of the AI Overview rollout that has already partly passed.

Future predictions

• Google's full-year 2026 ad revenue growth will likely land meaningfully above eMarketer's February forecast of 5.6%, given the Q2 print of 17% — expect a forecast revision before year-end (Confidence: Medium — based on the gap between forecast and actual results already observed).
• The February 2027 DSA-to-AI-Max migration deadline will face at least one more delay or partial carve-out, following the pattern already set by the June 2026 reopening of DSA campaign creation (Confidence: Low — extrapolated from one data point, not a confirmed trend).
• Paid CTR on AI-Overview-present queries will continue the partial-recovery trend Seer documented through February 2026, as advertisers adapt ad formats and bidding to AI-generated result pages (Confidence: Medium — based on two data points from the same research source).
• Amazon's share of US search ad spend will continue climbing toward or past the 25.7%-by-2027 eMarketer projection, pulling incremental retail-intent budget away from Google (Confidence: Medium — based on a directly sourced forecast, not yet realized).
• Expect continued divergence between Google's self-reported AI Max performance claims (27% more conversions, 6x return) and the agency-reported CPC increases documented above — this gap is unlikely to close without independent, third-party audited performance data, which does not yet exist (Confidence: Medium).
• WARC's global ad spend forecast for 2026 will likely be revised upward again before the year closes, continuing the pattern of positive revisions already seen twice (Confidence: Low — based on a short revision history, not a long-run pattern).

Practical takeaways

1. Stop budgeting off a single CPC benchmark. The $5.42 cross-industry average masks a 50-point swing between Real Estate (+27.27% YoY) and Education (-22.79% YoY) — pull your own vertical's number before setting expectations for next quarter's spend.
2. Treat AI Max adoption as a query-expansion decision, not a settings toggle. Because it opens auctions for queries you weren't previously bidding on, model the incremental spend before you model the incremental conversions — teams doing structured Technical Foundations work on their measurement stack will catch this gap earlier.
3. Don't plan around the 68% AI Overview CTR collapse figure — it's five months out of date. Use Seer's recovery trajectory (6.34% to 16.21%) as your baseline assumption and revisit quarterly.
4. Track your brand's actual presence inside AI Overviews and AI Mode answers, not just your paid rank — with AI Mode queries running 3x longer and doubling every quarter per Google's own numbers, LLM Visibility Monitoring is no longer optional context, it's a budget input.
5. Reallocate a fixed test budget toward retail media and non-search channels now, ahead of the curve eMarketer and Forrester are both describing, before CPC pain forces the decision on worse terms.
6. Build entity- and brand-level presence that doesn't depend on a single auction — Semantic Authority and Entity & Knowledge Architecture work compounds independently of what Google charges per click this quarter.