Google’s market share is roughly 74% — not 90%. That gap is where your next pipeline comes from.

Here’s a number that should recalibrate how B2B marketing teams allocate attention in 2025: Google controls 73.7% of all desktop searches across 41 major websites analyzed by Rand Fishkin and the teams at SparkToro and Datos. Not 90%. Not the dominant-to-the-point-of-irrelevance figure that most market share reports publish. Seventy-three percent — which means more than a quarter of all desktop search activity is happening somewhere else entirely.

That somewhere else is where this research gets interesting.

## The Study That Changed How to Think About Search

In March 2026, Fishkin published findings from an analysis of 41 websites with significant search activity, drawing on Datos’ 2025 desktop panel covering millions of devices across the United States, 27 EU member countries, and the United Kingdom. The study ran the full 2025 calendar year — January through December — and covered every major vertical: traditional search engines, e-commerce platforms, AI tools, social networks, reference sites, travel platforms, real estate, and classifieds.

The core argument is deceptively simple: search is a behavior, not a channel. People don’t just search on Google. They search on Amazon, YouTube, Reddit, LinkedIn, Pinterest, Threads, and dozens of other platforms — and those searches carry intent signals that most demand generation teams are completely ignoring.

For B2B marketers who’ve spent the last two years obsessing over ChatGPT’s rise, the data delivers a sharp correction.

## ChatGPT Is Smaller Than You Think

Among Q4 2025 desktop search activity across the 41 analyzed domains, AI tools collectively captured 3.2% of all searches. ChatGPT — despite the volume of industry conversation dedicated to it — trails Amazon, Bing, and YouTube in actual search volume. On desktop, where AI tools should theoretically be strongest, ChatGPT ranks behind Reddit, Amazon, Facebook, YouTube, and Google in raw visitor reach.

The engagement data makes this starker. Only about 50% of people who visit ChatGPT actually submit a prompt. The other half are viewing chats shared by others — passive consumption, not active search. Compare that to Google, where nearly 100% of visitors conduct a search, or Amazon and Bing, where more than 70% of visitors search.

This doesn’t mean AI search is irrelevant. It means the hype has outrun the data. Organizations reallocating meaningful budget toward ChatGPT visibility optimization while neglecting Amazon, YouTube, and Bing optimization may be chasing the story rather than the signal.

The research is clear on this point: if you’re concerned about AI search visibility, you should simultaneously be focused on search presence in Amazon, Bing, and YouTube — because those platforms receive more desktop search activity than ChatGPT, and your audience is likely already there.

## Where Search Share Actually Moved in 2025

The year-over-year data from 2025 tells a story that cuts against most conventional narratives about search consolidation.

Google lost 3.5 percentage points of US desktop search market share during 2025 — the only metric across major web behavior categories where Google showed decline. In the EU and UK, the loss was smaller (roughly 2 percentage points), but the direction was the same.

What grew? Amazon, Bing, and YouTube all increased their share. ChatGPT, despite its reputation as the category disruptor, actually decreased its share in the US and barely held flat in Europe.

More telling: 34 of the 41 sites outside the top 7 grew their share of search during 2025. Fishkin describes this as one of the only areas of web behavior in the past decade where the largest platforms did not become more dominant over time. Search is genuinely fragmenting — not toward AI tools specifically, but toward a broader ecosystem of platforms where people conduct queries.

For demand generation professionals, this fragmentation is both a problem and an opening. The problem: your audience’s search behavior is distributed across more surfaces than your current analytics stack probably captures. The opening: most of your competitors haven’t figured this out yet.

## The Platforms You’re Probably Undervaluing

Some of the most actionable findings in this research involve platforms that rarely appear in demand gen strategy decks.

**Threads** grew from approximately 2 searches per searcher per month in January 2025 to nearly 4 by December — a near-doubling of search intensity on a platform that most B2B marketers still treat as experimental. In January 2026, Threads surpassed Twitter/X in daily mobile users. The trajectory matters.

**Pinterest** showed significant search growth throughout 2025 and now sits in the 5–10 searches per searcher per month range — the same tier as eBay. For B2B companies in design-adjacent, creative, or visual industries, Pinterest search optimization is likely more productive than most teams realize.

**Instagram** made a notable shift: substantial rise in desktop traffic throughout 2025, transitioning from a mobile-first platform to one with meaningful desktop presence. That changes what optimization for Instagram actually requires.

**LinkedIn** surfaces a counterintuitive finding. Despite being the obvious home for professional people-search, LinkedIn’s native search generates only 1–5 searches per searcher per month — placing it in the lowest engagement tier of the study. The likely explanation: Google simply does LinkedIn search better than LinkedIn does. Users prefer to search LinkedIn content via Google rather than through LinkedIn’s own search bar. This has real implications for how B2B content teams should think about LinkedIn SEO versus LinkedIn-native discoverability.

**Reddit** shows a similar pattern — consistent traffic growth throughout 2025, but low on-platform search rates. Again, Google’s superior search of Reddit content means users are finding Reddit threads through Google, not through Reddit’s own search function. The implication for B2B content strategy: Reddit presence matters for Google rankings, not primarily for Reddit-native discovery.

## What This Means for Demand Generation Strategy

The research makes a recommendation that’s worth sitting with: SEO should evolve to mean Search Everywhere Optimization.

This isn’t a rebrand. It’s a structural shift in how demand generation teams define their job. The traditional model — optimize for Google, maybe Bing, increasingly worry about ChatGPT — accounts for roughly 80% of desktop search activity across traditional engines. But the remaining 20% is distributed across commerce sites (~10%), social networks (~5.5%), and AI tools (~3.2%), and within each of those categories, the specific platform distribution matters enormously.

For B2B specifically, several implications stand out.

First, the intent data sitting inside your own website search bar is almost certainly underutilized. When a prospect searches your site for “enterprise pricing” or “integration with Salesforce,” that’s a buying signal — not a UX event. Intent data tools like 6sense and Sona can now tie anonymous site search behavior to specific accounts and buying stages, transforming what used to be a navigation feature into a revenue intelligence feed. Most B2B teams haven’t made this connection.

Second, the distinction between visits and searches matters more than most analytics setups reflect. The research found that visits are not a reliable proxy for search engagement — a finding that applies directly to how B2B teams evaluate platform performance. A platform with high traffic but low search engagement is a different asset than one with lower traffic and high search intensity. Optimizing for the wrong metric produces the wrong conclusions.

Third, the monetization model for search is shifting. Revenue teams should note that the economics are moving from referral clicks toward API access, sponsored data feeds, and co-branded conversational experiences. This changes how search-driven demand generation gets measured — and how it gets valued internally.

## Google’s Real Position — and What AI Actually Changed

One of the study’s most important contributions is correcting the narrative around Google’s decline.

Google lost market share in 2025. That’s real. But it remains the dominant search platform by a significant margin — and here’s the part that surprises most people: Google is also the largest AI search tool. Approximately 16% of Google SERPs show AI Overviews. Less than 0.1% of searchers click to AI Mode — but AI Overviews are appearing at scale. Even combining all prompts from ChatGPT, Claude, Deepseek, and every other AI tool, Google’s AI search volume exceeds them all combined by at least an order of magnitude.

The practical implication: the shift to AI-powered search is largely happening inside Google, not instead of Google. For demand generation teams, this means structured, crawlable, authoritative content remains the primary currency for search visibility — across both traditional results and AI-generated answers. Generative Engine Optimization layers on top of SEO; it doesn’t replace it.

Only 26% of organizations have scaled AI to sustained value, according to McKinsey. That’s a caution worth carrying into any conversation about reallocating resources toward AI search optimization. The infrastructure investment required — vector databases grew 377% year-over-year, reflecting how seriously enterprises are taking AI-powered search — is real, and the ROI timeline is not yet clear for most organizations.

## The Research Gap That Matters

This study covers desktop behavior only. Mobile browsers and apps — where a substantial share of total search volume occurs — are excluded by design. The 41 domains were editorially selected by Fishkin from the most-visited 250 domains in Datos’ panel, which means thousands of domains with search functionality weren’t counted. The methodology treats all searches equally: a complex multi-turn ChatGPT session counts the same as a single Google query with zero clicks.

These aren’t criticisms. They’re honest constraints that the research acknowledges directly. But they mean the actual fragmentation of search behavior is likely even more pronounced than the data shows. If 23 of 41 analyzed sites had more than 0.1% market share of desktop search — and assuming Google’s 2025 search volume matched 2024’s approximately 5 trillion searches, that 0.1% threshold represents 5 billion searches — the long tail of search activity across the full web is enormous.

For demand generation teams, the honest takeaway is that the map you’re using to navigate search behavior was drawn for a simpler terrain. The actual landscape has more surfaces, more platforms, and more intent signals than the standard Google-plus-ChatGPT framework accounts for.

The teams that build measurement and optimization systems reflecting this complexity — not the ones that wait for the narrative to settle — are the ones who’ll find pipeline in places their competitors haven’t looked.

Search happens everywhere. The question is whether your demand generation strategy is built to see it.