Stratrix

A million visits a week

On October 29, 2026, Reddit told its shareholders something every company that lives on search traffic has been dreading. “Our search referral traffic was down meaningfully in Q3, with the decline accelerating later in the quarter,” its letter said. October was worse: “search traffic has dropped by more than 1 million referrals on average per week.” If that went on, the company warned, its daily users could fall in Q4 “by a low-double-digit percentage” from about 130 million.

The next day Reddit’s shares fell 29%, their worst day since the company went public in 2024. That is the market’s business, not ours. Ours is the question underneath it, and it applies to every company that finds customers through search.

That AI answers take clicks from the websites they summarize is not news. Anyone who has used Google lately has seen the answer appear above the links. The more useful questions are what the companies that have lived with it longest are doing about it, and where the money in search is going instead. Their filings answer both, and the answer is not to fight for the click.

Here is where we stand, and why.

Where we stand

Through 2027, search will send fewer free visitors to the sites its AI answers are built from, and Google will keep more of the value on its own pages. The businesses that hold up will stop measuring success in visits and turn each visitor they still get into someone they know: a member, a subscriber, an app user, a licensee.

No longer a few complaints

We searched every annual report, quarterly report and earnings release filed with the SEC since January 2023 for the language of AI in search: AI Overviews, zero-click, answer engines and their cousins. Then we kept only the sentences that also talk about traffic.

The count went from 7 companies in 2024 to 24 in 2025 and 52 in the first ten months of 2026. Most still hedge: in 2026, 21 of the 52 only warn that it may hurt them. But 8 now say it already has.

Companies whose SEC filings tie AI answers in search to their traffic
YearCompaniesSay it has already happened
20247—
202524—
2026, January to October528

Distinct companies with at least one sentence in a 10-K, 10-Q or 8-K naming AI in search (AI Overviews, zero-click, answer engines and similar) together with traffic. The last column is our hand classification for 2026; 21 more only warn that it may happen. Source: SEC EDGAR full-text search; Stratrix classification.

Their sentences are blunt. Chegg, the homework help company, lost 40% of its subscribers in a year “as we continue to feel the impact of lower traffic, largely due to Google AI Overviews.” People Inc, the publisher behind PEOPLE, reported a 22% decline in its core sessions in the second quarter of 2026, “due primarily to the impact of the increasing prominence of Google AI Overviews on Google search sessions.” LegalZoom says the shift toward zero-click searches “has contributed, and may continue to contribute, to reduced organic traffic, increased customer acquisition costs” and weaker search marketing.

The toll road

Now look at the other side of the search page. Every quarter Alphabet reports two numbers that, side by side, tell a story. One is paid clicks: clicks on the ads Google shows on its own search pages and properties. The other is impressions on the Google Network: the ads Google places on other companies’ websites and apps.

Paid clicks have grown in every year since 2023, and by 13%, 13% and 8% from a year earlier in the first three quarters of 2026. Network impressions have fallen in every year over the same stretch, and by 14% in the third quarter of 2026.

Google’s paid clicks and the ads it places on other sites, change from a year earlier
−20%−10%0%10%20%Paid clicks on Google’s own pages, 2023: 7%Impressions on ads Google places on other sites, 2023: −5%2023Paid clicks on Google’s own pages, 2024: 5%Impressions on ads Google places on other sites, 2024: −11%2024Paid clicks on Google’s own pages, 2025: 6%Impressions on ads Google places on other sites, 2025: −7%2025Paid clicks on Google’s own pages, Q1 2026: 13%Impressions on ads Google places on other sites, Q1 2026: −9%Q1 2026Paid clicks on Google’s own pages, Q2 2026: 13%Impressions on ads Google places on other sites, Q2 2026: −12%Q2 2026Paid clicks on Google’s own pages, Q3 2026: 8%Impressions on ads Google places on other sites, Q3 2026: −14%Q3 2026
  • Paid clicks on Google’s own pages
  • Impressions on ads Google places on other sites
Show the numbers
Paid clicks on Google’s own pagesImpressions on ads Google places on other sites
20237%−5%
20245%−11%
20256%−7%
Q1 202613%−9%
Q2 202613%−12%
Q3 20268%−14%

Paid clicks are clicks on ads on Google Search and other Google properties. Google Network impressions are ads on partner sites and apps. Alphabet does not attribute either trend to AI answers. Source: Alphabet Inc., Forms 10-K and 10-Q.

Alphabet does not say AI answers caused either trend, and both have many causes. But together they describe a toll road. More of the attention search creates stays on Google’s own pages, where Google sells it. Less of it flows out to the rest of the web, where everyone else used to earn from it. For a business that used to get the click for free, the choice is to buy it back as an ad or to do without it. LegalZoom’s higher customer acquisition costs are what buying it back looks like.

Losing the visit is not losing the business

If the story ended there, it would be a story about decline. People Inc’s numbers say it does not have to end there. In the first quarter of 2024, Google search sent it 1,245 million sessions, 55% of its core sessions. Two years later it sent 463 million, 25%. That is the 63% decline in Google search referrals the company reports “over two years.”

Over the same two years its quarterly digital revenue rose from $221 million to $253 million. Two things carried it. Revenue from the sessions it still had held flat, at $149 million and then $150 million, on 19% fewer sessions: about 24% more revenue per session, by our arithmetic. And revenue that does not depend on sessions at all, from AI licensing, apps, social platforms, events and commerce, grew 43%, to 41% of the total. Views of its content away from its own sites roughly doubled, from 10,192 million to 20,962 million.

People Inc: Google visits fell, revenue grew
Q1 2024Q1 2025Q1 2026
Core sessions from Google search (millions)1,245759463
Core sessions from everything else (millions)1,0271,4521,378
Views of its content on other platforms (millions)10,19216,54520,962
Revenue from sessions ($ millions)$149$152$150
Revenue not from sessions ($ millions)$72$83$103
Digital revenue ($ millions)$221$235$253

As reported in the company's presentation. Core sessions cover its core brands; revenue not from sessions includes AI licensing, apps, social platforms, events and commerce. Source: People Inc., Q1 2026 earnings presentation (Form 8-K, Exhibit 99.2).

AI Overviews now appear on nearly 70% of the company’s top search queries, by its own count. It stopped treating the Google visit as the business, and the business grew.

A known visitor is worth more

Reddit’s quarter, under the headline, says something similar. Its daily users in the United States fell 2% from a year earlier and 5% from the quarter before. Yet its revenue per daily user rose 36%, to $6.87, mainly on higher ad prices.

Its answer to the search decline is to stop serving strangers. Visitors who arrive from search engines, its quarterly report notes, “are generally not logged in.” The company is now “shifting Reddit to a logged-in-only experience,” because, its letter says, “we will have more signals for ads targeting, and being logged-in leads to higher engagement generally.”

Other companies are finding the same thing from other directions. TechTarget, which sells to technology buyers, reports “a 2x to 3x higher membership conversion rate from answer engine and LLM citations compared to traditional organic search”: the people who come through AI answers arrive readier to sign up. Etsy says clicks from its own email and push messages rose more than 25% in the fourth quarter of 2025.

The best objections

Three deserve an answer.

First, walls cost visitors. Reddit says so itself: its experiments with requiring a login “could have adverse impacts on our business.” Some people who arrive from search will leave rather than sign up. That is fair, and it is why we would not copy Reddit’s move wholesale. The point is not the wall. It is knowing who your visitors are, and a wall is only one way to get there.

Second, People Inc is one company, with famous brands and content that AI companies will pay to license. A small business cannot replace search with licensing deals and Apple News. Also true. But the idea scales down: a mailing list, a customer account, a reason to come back directly. Etsy is not a publisher, and its own channels grew.

Third, the ads Google places on other sites could be falling for reasons that have nothing to do with AI answers, and Alphabet does not say. That is why we will keep watching that number rather than assume it, and why a turn in it would change our mind.

What we think

For decades the deal was simple. You published, search sent visitors, and the visit was the business: an ad view, a lead, a sale. That deal is ending, not because people search less, but because the search page has learned to answer the question itself and keep the attention it creates.

We think the right response is not to fight for the old click. It is to change what you count. A visit you cannot identify is now a lead, not an audience. The companies holding up measure the people they know, members, subscribers, app users, licensees, and how much each one is worth. They treat search, paid or free, as one way among several to meet them.

The window matters. Search still sends real traffic today, and every visitor who arrives is a chance to start a relationship you will still have when the free click is gone. Businesses that use the next two years to turn search visitors into known customers will feel the rest of the decline far less. Businesses that keep optimizing for the visit will keep paying more for less of it.

What to do about it

CEO and board

  • Stop accepting traffic as the headline number. Ask for the number of customers the business can reach directly, and what each is worth.
  • Ask what share of new customers still arrives through free search, and what it would cost to replace them with paid clicks.

CMO and growth

  • Give every search visitor a reason to become known on the first visit: an account, a newsletter, an app, a saved list.
  • Track visitors who arrive from AI answers separately; they come with more intent and may deserve different treatment.
  • Before putting up a login wall, test what it costs in visitors, as Reddit is doing.

CFO

  • Price customer acquisition as if free search traffic keeps falling, and treat what remains as a bonus.
  • Judge revenue that does not depend on visits, such as licensing and direct audiences, separately from revenue that does.

CIO and COO

  • Make your content easy for AI answers to cite and link to, and measure those citations.
  • Build the customer accounts and data that make known visitors possible.

What would change our mind

We could be wrong. These are the signs we are watching, what each would tell us, and where each stands today.

If People Inc's visits from Google stopped falling from a year earlier, the slide in free search visits would have bottomed out for one of the publishers most exposed to it, and the urgency here would fade. Today: 39%, as of March 31, 2026.

If the ads Google places on other companies' sites started growing again, more of the money search makes would be flowing back out to the rest of the web, and the toll-road picture would be wrong. Today: 14%, as of September 30, 2026.

If People Inc's digital revenue started shrinking from a year earlier, the best example of losing Google visits and growing anyway would have stopped working. Today: 8%, as of March 31, 2026.

If Reddit's revenue per daily user grew less than 10% a year while it moves to logged-in users, knowing your visitors would be worth less than we think. Today: 36%, as of September 30, 2026.

None of these has moved far enough to change our mind.

How sure we are

Fairly sure. The evidence is three years of securities filings from both sides of the search page: the sites losing visits and the company keeping them. But no independent measure of web traffic is in it yet, People Inc is a single example of losing Google and growing anyway, and Alphabet does not say why the ads it places on other sites keep falling. We are surer that the free visit is shrinking than of how easily others can copy People Inc.

How we judge confidence

Every call gets the same five checks, each scored 0, 1 or 2, so confidence is worked out rather than felt. Up to 4 of 10 is low, 5 to 7 is medium, and 8 or more is high, as long as no check scores zero. This call scores 7 of 10, which is medium.

  • Source, 2 of 2. Every figure comes from securities filings: Reddit's, Alphabet's, People Inc's and those of 52 other companies, with one news report for the share move.
  • Breadth, 1 of 2. Filings from both sides of the search page point the same way, the sites losing visits and the company keeping them. But they are all company disclosures; no independent measure of web traffic is in the evidence yet.
  • Duration, 2 of 2. Alphabet's click and impression figures cover every year from 2023 and every quarter of 2026, the disclosure count covers three years, and People Inc's figures cover two years.
  • Counter-evidence, 1 of 2. Login walls cost visitors, People Inc is one company with unusual assets, and the ads Google places on other sites may be falling for other reasons. All three are answered on this page, in The best objections.
  • Precedent, 1 of 2. Not yet tested against our Decision Records on earlier platform shifts, such as publishers who lost social media referral traffic. Scored as untested until it is.

We will grade ourselves

On November 3, 2027, a year after publishing, we will come back to this page and say whether we were right, and whether for the reasons we gave. This page will stay as written, with the review added below it.

Sources

  1. Reddit, Inc., Q3 2026 Letter to Shareholders (Form 8-K, Exhibit 99.2, October 29, 2026)
  2. Reddit, Inc., Form 10-Q for the quarter ended September 30, 2026
  3. CNBC, Reddit third-quarter 2026 earnings and share move (October 30, 2026)
  4. Alphabet Inc., Form 10-Q for the quarter ended September 30, 2026
  5. Alphabet Inc., Form 10-K for 2025
  6. People Inc., Q1 2026 earnings presentation (Form 8-K, Exhibit 99.2)
  7. People Inc., Form 10-Q for the quarter ended June 30, 2026
  8. Chegg, Inc., Q2 2025 results (Form 8-K, Exhibit 99.01)
  9. LegalZoom.com, Inc., Form 10-Q for the quarter ended June 30, 2026
  10. TechTarget, Inc., Form 10-K for 2025
  11. Etsy, Inc., Q4 2025 results (Form 8-K, Exhibit 99.1)
  12. SEC EDGAR full-text search, Forms 10-K, 10-Q and 8-K, January 2023 to October 2026

Not investment advice. Stratrix Stance does not recommend buying or selling any security; Reddit's share move is mentioned only as the news that prompted this call. Figures come from company filings as cited. The levels that would change our mind are Stratrix's judgment, not research.

Tools to try

More from Stratrix

Strategy Packs
The Road Not Taken

“What did the decision you did not take actually cost?”

One decision per pack: the framework, a working model, the sourced cases and the board material.

See the pack — $499 →

Weekly email

Stratonomics, every Tuesday

One data story every Tuesday: what happened, what it means, and the figures you can quote. Free.

We'll email you a link to confirm. Unsubscribe anytime. Privacy policy.

The Business Model Wizard

Make the seven choices that define your model, test your numbers against real companies, and meet the rival AI would build against you.

Free · about twelve minutesBuild your Model Thesis →