Fundamentals · 27 October 2026

By Eddie Eastwood, Founder

Where Your Buyers Actually Search: Google, AI and Beyond

How Big Is the Gap Between Google and ChatGPT, Really?

There is no clean number for how much bigger Google is than ChatGPT: Google has never published a total count of how many searches it handles. ChatGPT reports its own scale in messages, not searches. By July 2025, 700 million weekly active users on ChatGPT's consumer plans were sending 18 billion messages a week, more than 2.5 billion a day, according to NBER working paper 34255, "How People Use ChatGPT." Google, meanwhile, states that its total search query volume reached an all-time high last quarter, and that AI Mode, a feature inside Search, passed 1 billion monthly users within a year of launch, with its own query count more than doubling every quarter. None of those three figures can be subtracted from one another to produce a single "Google is N times the size of ChatGPT" answer, because none of them are measuring the same thing.

Google's developer-facing "How Search Works" page states only that "billions of people use Search," with no count attached. Alphabet's own earnings filings for the past two quarters describe Search query growth only in words, never in a number. The Q2 2026 exhibit states "Our popular AI features are driving Search query growth" in the same breath that Gemini models are reported to "process 22 billion API tokens per minute" with "the Gemini App has 950 million monthly active users." The Q4 2025 exhibit pairs "Search saw more usage than ever before, with AI continuing to drive an expansionary moment" with Gemini processing "over 10 billion tokens per minute" and the Gemini App reaching "over 750 million monthly active users." Tokens and Gemini App users get a precise count in both filings; Search's own query growth does not. Sundar Pichai's I/O 2026 keynote reports that Google now processes 3.2 quadrillion tokens a month, up sevenfold from 9.7 trillion two years earlier. That is a real number about a real jump in AI compute. It is not a search count. A token is a fragment of text an AI model processes, generated in far larger quantities per query than a search produces, so converting it into an implied number of searches would be inventing a figure Google never gave. The widely repeated "8.5 billion searches a day" and "5 trillion searches a year" figures have no published Google source in the evidence available here. They cannot establish a total-query comparison. We have run down the other AI-search figures that fail a primary-source check in our review of the state of AI search in 2026.

What can be compared, on one consistent basis, is monthly reach. Search engines like Google and Yahoo are used by 80.3% of online adults within a given month, the lowest that figure has ever measured in GWI's ongoing tracking survey, cited in DataReportal's Digital 2026 Global Overview Report. LLM and GenAI platforms, across the same report, reach an estimated more than 1 billion people monthly, a figure DataReportal itself flags as an extrapolation across multiple data sources rather than a single measured count. These public figures describe different channel-level measures. They should not be converted into a Google-versus-assistant comparison or a search-to-LLM ratio. Both channels are growing from their own base: Google's AI Mode queries are doubling every quarter, and ChatGPT grew from roughly 100 million weekly users after one year to almost 350 million after two, reaching 700 million by its third.

Google's total site visits, not its query volume, are what is actually shrinking here. Kepios's rolling eight-week analysis found that Google's total site visits have "actually declined by somewhere between 1 and 2 percent over the past 2 years," even as Google still accounts for close to 90% of all search engine referrals to third-party sites, per Statcounter data in the same report. We go deeper on exactly how that referral mechanic plays out between ChatGPT and Google in our comparison of ChatGPT and Google organic traffic, and on what happens to the click itself when an AI-generated answer appears above the result in our piece on the zero-click search era. DataReportal's own reading of its data is that Google is converting fewer of its searches into referral clicks, which is a more specific and more useful description than "Google is losing to AI."

Reach, queries, messages, referrals and users are five different measurements, and the table above exists because conflating any two of them is the single easiest way to get this topic wrong. No clean comparison number exists for Google versus ChatGPT, for the reasons above: one side does not publish a query count, and the other reports messages, not searches.

Why Showing Up Cannot Mean Google Alone

In its most recent SEC filing, Pinterest describes itself as "a visual search and discovery platform, positioned at the intersection of search, social and commerce," and reports 640 million monthly active users, up 11% year over year. That makes it a substantial place for visual browsing, product discovery and commercial research.

YouTube reports "over 20 million videos uploaded daily," YouTube Shorts "averaging over 200 billion daily views," and "billions of monthly logged-in users." Its published monthly-user wording remains broad. Separately, Nielsen data as of January 2026 puts YouTube in the number one spot for US streaming watchtime for nearly three years, a result limited to the United States and that period.

GWI's research, also in the DataReportal report, puts YouTube's monthly usage at 55.4% of online adults, a figure GWI classifies under video rather than social platforms, so it uses a different basis from the social-platform figures elsewhere in the report and is not directly comparable to Google's 80.3% search-engine reach. The verified figures point to uploads, Shorts viewing and US streaming watchtime; they do not establish a global ranking among search engines.

A business that appears only in Google's results leaves no direct evidence for people researching visually on Pinterest or watching demonstrations on YouTube. Those are distinct discovery habits, not substitutes for a traditional search box.

How Do Buyers Actually Research a Purchase?

The average online adult uses 6.75 different social media platforms every month, according to GWI's latest wave of research in the same DataReportal report. If a customer's attention is already split across nearly seven platforms before they ever reach your website, "showing up" cannot mean occupying one of them and assuming the rest take care of themselves.

Search still wins the discovery contest outright. GWI's research finds that search engines remain the number one source of brand discovery worldwide, with nearly one in three internet users aged 16 and over saying they learn about new products and services through platforms like Google and Bing, and that search engines are also the primary destination people turn to when actively researching a purchase. Social media advertising sits in third place overall behind search engines and TV ads, with more than 30% of adult internet users saying they discover brands through ads on social platforms, but for the 16 to 34 age group specifically, social ads are the single top driver of brand awareness, ahead of search. LinkedIn, TikTok and Facebook sit alongside YouTube and Pinterest as places a buyer might encounter a brand before they ever type a query into Google.

That age split describes one buyer population moving through a discovery-to-evaluation loop on different default channels depending on age. It gives a business a reason to check which channels its own customers use before setting a visibility plan.

These platforms carry different weight in that habit, measured on GWI's social-platform basis, which differs from the search-engine reach figure above and is not comparable to it. Facebook has "the largest audience of online adults aged 16 and above," with 56.9% of respondents in the Q2 2025 survey saying they used it "at least once within the previous 30 days." GWI also finds that users 65 and above are the most likely to use Facebook, even as the platform stays "consistently popular across all other age groups too." TikTok ranks sixth among social platforms at 36%, a figure that excludes China because GWI's China survey option refers to Douyin, a separate product. Pinterest, beyond its scale as a discovery platform covered above, has its strongest reported audience skew among female internet users aged 16 to 24 rather than any other demographic. That matters for a business targeting that group specifically.

LinkedIn describes itself as "the world's largest professional network with more than 1 billion members in more than 200 countries and territories." That is a cumulative membership count rather than a monthly-use figure. For B2B firms, it makes LinkedIn a sensible place to show named people, expertise and company context when those are the proof a prospect seeks.

LinkedIn, TikTok and Facebook can contribute different evidence to an evaluation. Google's "messy middle" research describes exploration and evaluation occurring "across a huge array of online sources, such as search engines, social media, aggregators, and review websites." A LinkedIn profile can establish professional credentials; a Facebook group can reveal local discussion; a review site can expose service experience. Each action answers a different part of a buyer's decision.

Google's consumer research into the "messy middle" of the purchase journey, published by Think with Google in July 2020, found that the path between first noticing a need and actually buying "is not linear" and involves "a complicated web of touchpoints that differs from person to person." The same research describes two repeating mental modes, exploring options and evaluating them, with people cycling between the two "as many times as they need to make a purchase decision." In a large-scale experiment simulating 310,000 purchase scenarios across financial services, retail, travel and other categories, even the weakest challenger brand tested, a fictional cereal brand, still won 28% of shopper preference away from the established market leader once it was "supercharged" with stronger reviews and a better offer. That result is specific to a 2020 study and a simulated scenario, not a universal law, but it demonstrates that stronger reviews and an improved offer can shift preference in a tested setting.

Nobody, including Google's own research, puts a number on exactly how many times a buyer needs to encounter a brand before they choose it. The "seven touchpoints" rule has no primary source behind it, and the Think with Google research that describes the loop assigns no count. It supports planning for several moments of exploration and evaluation, rather than treating one appearance as a reliable route to a sale.

How Should a Business Map Its Channels to Buyer Research?

Start with one question a real prospect must answer. For a local B2B provider, it may be: Can this company solve my problem in my area, and can I trust the people who will carry out the maintenance? Put the basic answer on the company website, the relevant service page and its Google Business Profile: the service offered, geography served, a named contact route and evidence that the business operates there. That gives a person who searches the category a way to shortlist the firm.

Then match the proof to the part of the decision that remains. A buyer who needs to see how the service works can use a YouTube demonstration or case-study video. A buyer checking professional competence can find the named specialists, company description and point of view on LinkedIn. A buyer whose choice is visual can use Pinterest to inspect examples. A buyer who needs local reassurance may consult a Facebook group, marketplace listing or local discussion. There is no reason to publish the same material on all of them: choose the two or three places that answer the questions your buyers actually ask.

The next research step follows from what is missing. If the service page earns visits but prospects ask whether the firm has experience in their sector, publish a named case study and link it from the relevant LinkedIn profile or video. If local prospects ask whether you cover their postcode, make that geography explicit on the website and Google Business Profile before creating another social post. This turns channel selection into a sequence of buyer decisions rather than a posting schedule.

A prospect comparing the service page with a LinkedIn profile or video should find the same basic business, then a different kind of proof. Select the evidence a buyer needs at each step, then keep the underlying business facts consistent.

Consider an illustrative route for a facilities manager who needs a provider to solve a particular maintenance problem at premises in her area. Her question is not whether the provider can publish a persuasive post. It is whether the company handles this exact kind of problem locally, who will take responsibility for the enquiry, and whether the claimed method makes sense before she puts the firm on a shortlist. That question gives the research a practical order. It also limits the places the business needs to maintain well.

She can begin with the relevant website service page and the Google Business Profile. The service page should state the scope of the offer in terms she recognises, make the service area plain and offer a direct way to make contact. The Business Profile should support the same basic account of what the company does and where it operates. They do not need to settle every concern, because their job in this example is to establish that the provider is relevant and reachable.

If the manager needs to understand how the provider handles the maintenance task, a demonstration video can show the method, the equipment or the sequence of a typical visit. If she needs to assess competence, a named professional profile or relevant case material can identify the people involved, their experience and the kind of outcome they have handled. If local circumstances matter, she can look at a community discussion, local group or other local reference for lived context about access, timing or the type of premises involved.

The business has a corresponding discipline. Its company description, offer, named people, geography and contact route must agree wherever this route sends the manager. A video should not describe a different service area from the service page. A professional profile should not introduce a specialist who cannot be connected to the company contact route. Case material should make clear which service it illustrates. Consistency lets each item add its own proof without forcing the prospective customer to work out whether she has found different versions of the same provider.

Conflicting information changes the next action. If the Business Profile names a wider area than the service page, correct the weaker or outdated description before asking people to trust a new campaign. If the case material suggests a service that the current offer no longer includes, revise the case introduction or remove it from the route. If the community discussion exposes a practical concern that the owned pages do not answer, add a clear explanation where the buyer first needs it and make the contact route able to handle the question.

What Should You Actually Do About It?

Google's recommendation, drawn from the same messy middle research, is to "ensure brand presence so your product or service is strategically front of mind while your customers explore." Google's total queries are at an all-time high by its own account, and it still carries close to 90% of all search referral traffic. Search therefore remains the core place to make the offer, location, proof and contact route easy to find.

Add channel-specific proof where it helps a buyer continue the research:

• Keep your Google Business Profile and website accurate and complete, especially the service description, geography, evidence and contact route.
• Put demonstrations or case-study footage on YouTube when a buyer needs to see the service in use; use Pinterest when visual examples drive the decision.
• Use LinkedIn to show the people, credentials and specialist perspective behind a B2B offer. Use Facebook where a local group, marketplace or community conversation affects trust. Use TikTok or X only when their formats and audiences suit a specific buyer question.
• Check that reviews, comparisons, case studies and third-party mentions describe the same company and offer as your owned profiles.

The point is selective coverage, not daily posting across six platforms. Search retains its central role, while the supporting channels carry the proof that a particular buyer may seek before making contact. Consistent, useful evidence across the places your customers actually consult gives them fewer gaps to resolve as they move from first search to a shortlist and a decision.

Google remains a core part of buyer research: its total query volume is at an all-time high, and it accounts for close to 90% of search-engine referrals to third-party sites. Search-engine monthly reach and estimated monthly LLM/GenAI-platform users are different channel-level measures, so they cannot produce a Google-versus-assistant ratio. The practical planning task is to keep the offer, geography, people and contact route clear in search, then add the demonstrations, credentials, visual examples or local proof that buyers seek on the limited channels relevant to their decision. That is how a business supports a research path that runs across more than one destination.

Frequently asked questions

Is Google still bigger than ChatGPT?

There is no single verified number that settles it either way, because Google does not publish a total count of searches handled and ChatGPT reports its own scale in messages, not searches. Search engines reach 80.3% of online adults monthly, while LLM and GenAI platforms reach an estimated 1 billion or more. Those public figures describe different channel-level measures, so they should not be converted into a Google-versus-assistant reach claim or a ratio. On growth, both are expanding: Google's total queries hit an all-time high last quarter, and ChatGPT grew from roughly 100 million weekly users after one year to 700 million by July 2025.

Is Pinterest a search engine?

Pinterest describes itself that way in its own SEC filing, calling itself "a visual search and discovery platform, positioned at the intersection of search, social and commerce." It reported 640 million monthly active users as of June 2026, an 11% increase year over year. Whatever label you use, it functions as a discovery channel buyers actually use before they decide what to buy.

How many platforms do buyers use before they choose?

The average online adult uses 6.75 different social media platforms a month, according to GWI's research cited in DataReportal's Digital 2026 report. Google's own consumer research describes the buying process as a non-linear loop between exploring options and evaluating them, repeated across search engines, social media, aggregators and review sites until the buyer is ready to commit. No verified study puts a fixed number of exposures on that loop.

Should I move budget from Google to AI?

Keep search funded while testing where your buyers seek additional proof. Google's total queries are at an all-time high by its own account, and it still receives close to 90% of all search engine referral traffic. Google's total site visits are down 1 to 2% over the past two years, while LLM and GenAI platforms have grown into a separate channel with more than 1 billion estimated monthly users. The figures describe different measures, so they do not set a transfer formula for budget.

Does posting on social media help AI visibility?

It can support the broader buyer-research path when a post gives a prospect useful proof, such as a demonstration, specialist expertise or local context. Treat that as a visibility and evidence decision for the channels your buyers use. It does not establish that any single post changes an AI answer.

See exactly where your business shows up, and where it is still invisible, across the platforms your buyers actually use this month. QBiz Leads AI runs a free AI visibility check across search, AI assistants and the discovery platforms covered in this article, and hands you back a plain list of where to act first.

Run a free AI visibility check →