pressfit.aiResearch · November 2026

The Zero-Click
Mystery, Solved.

SparkToro found visits per internet user down 29.2% and couldn't say why. We lined its data up against jobs, prices, retail, apps and our own server logs.

SparkToro
"We can't say for certain."
The data
We can. Here's what happened.
The denominatorSparkToro, Similarweb, DataReportal

Visits barely fell. The user count exploded.

Most of the 29.2% headline is math: modestly fewer visits spread across far more people, much of them app-first users a browser panel can't count.

−12.2%
Top-million visits
÷
+24%
Internet users
=
−29.2%
Visits per user

Internet users, billions

DataReportal. The 2025 step is mostly an India and China data revision.

Share of visits by site rank

Similarweb via SparkToro. Big sites gained about 3 points. SparkToro's word: "stagnation."
Jobs and pricesChallenger, EIA, BLS

The people who browse all day lost their desks.

Office workers and household shoppers drive the open web. From 2023 on, layoffs hit the first and prices hit the second.

3.3M
U.S. job cuts announced, Jan 2023 to Oct 2026
Challenger, Gray & Christmas
628K
of them in tech, the sector that lives in a browser
168K, 134K, 154K, 172K by year
+26%
food-at-home prices, with shelter up about 28%
BLS, Dec 2020 to Dec 2025

Regular gas, annual average

EIA. Topped $5 in June 2022.

Announced job cuts, thousands

All sectors vs. technology. 2026 through October.
Retail splitCompany filings, Placer.ai

Shoppers stopped browsing. They consolidated.

Fiscal 2025 U.S. comparable sales. Costco grew four years running. Target fell in three of four, and its visits dropped after its January 2025 DEI rollback announcement. Analysts still debate how much was the boycott versus the economy.

Costco
+7.3%
93.2% renewal · 81M members
Target
−2.6%
Transactions −2.2% · visits −9% in Feb 2025

U.S. comparable sales by fiscal year

Costco excludes gas and FX. Different fiscal calendars.

Store visits, year over year, 2025

Placer.ai. Target fell seven straight months through August.
Apps and creatorsMeta, Similarweb, Substack, Semrush

Discovery moved where browser panels can't see.

Threads communities like Black Threads and creators like Ms Rachel send buyers straight from the app to small shops far below the top-million cutoff.

500M+
Threads monthly users
147M
Threads daily users
5M
Substack paid subscriptions
4.4x
value of an AI search visitor (Semrush model)
MeasurementPressfit.ai benchmark and telemetry

"Can't say for certain" is a tooling problem.

Ad-hoc prompts give different answers on the same input, and GA can't see AI traffic. Freeze the baseline, isolate each call, measure at the server.

Passed 4 runs in a row, 300-step pipeline

Share of 40 runs passing every deterministic check.

30 days on pressfit.ai

What Google Analytics saw vs. what our firewall saw.
99.1%
same answer twice, Opus 5.5
13.6%
verdicts flipped by the old batched step
−96.7%
judging cost, Haiku 5.5 vs. Opus 5.5
184
live ChatGPT and Claude fetches GA never saw

Stop guessing why traffic moved.
Engineer the proof.

The Pressfit Playbook shows where your competitors show up for your customers across Google, Gemini, ChatGPT, Claude and Perplexity, gives you prioritized steps to beat them, then measures it at the server.

Pressfit.ai Research · Sources: SparkToro and Similarweb (Nov 2026), DataReportal, PPC Land, Challenger Gray & Christmas, EIA, BLS CPI, Costco and Target filings, Placer.ai, Meta, Similarweb, Substack, Semrush, Pressfit.ai benchmarks and firewall telemetry · redevteam@pressfit.ai