
TLDR
Browser visits to the world's top million websites were 12.2% below their early-2022 peak by the third quarter of 2026, per SparkToro analysis of Similarweb data. The biggest sites did not absorb a larger share of the smaller pool. Australian marketers benchmarking against 2022 or 2023 figures are measuring against a pie that no longer exists.
KEY TAKEAWAYS
The numbers, and the correction
Rand Fishkin, co-founder and chief executive of SparkToro, published his analysis on 10 November 2026 using Similarweb's global desktop and mobile browser panel. Visits to the top million websites peaked at 2.243 trillion in the first quarter of 2022 and sat 12.2% below that level by the third quarter of 2026.[1]
One number needs careful handling before anyone puts this in a client deck. The URL slug still reads "24-in-5-years" while the live headline now reads 12%, which is citable evidence the post was corrected after publication.[1] Fishkin said on X on 10 November that the earlier version “had an incorrect graph/data”, Search Engine Journal reported.[2]
A second metric in the post, visits per internet-connected person down 29.2%, depends on DataReportal’s estimates of global internet users, which were revised partway through the period, so treat it as directional rather than precise.[2] The 12.2% headline drop is the more defensible figure.
Stagnation, not collapse
Fishkin is deliberate about the shape of the decline. Traffic held roughly level through 2023, dipped early 2024, and fell again in the first half of 2025. The top 100 sites held roughly the same share of visits throughout. Fishkin had expected Big Tech to be hoarding traffic; instead, he wrote, “stagnation is a more accurate description of the trend”.[1]
Fishkin also said total web traffic and visits per internet-connected person are both down from their 2021/22 peak, and he offered a list of possible contributors: app-only use, answers delivered on the search results page, AI tools, social platforms and the unwinding of pandemic-era browsing habits.[1] He did not identify a single cause, so anyone citing this research as proof that Google's AI Overviews killed web traffic is going further than the data supports.
What the data covers and excludes
Similarweb’s panel counts visits in desktop and mobile browsers. Activity inside apps is not counted.[2] There is no regional breakdown, so the global figure cannot serve as an Australian benchmark without local data alongside it.
That exclusion carries weight. Fishkin wrote that domination by Big Tech is likely far more prevalent in apps, and that traffic to US, UK and EU sites may have fallen while regions with growing internet access grew. Search Engine Journal’s advice is to treat the figures as global context, not a benchmark for any single market.[2]
The Australian read
For Australian publishers and marketers who set targets against 2022 or 2023 traffic, the finding changes the baseline. If the global pool has shrunk 12.2% since Q1 2022, any publisher holding flat may already be gaining share; it just does not show up against an outdated baseline.
The same logic applies to marketers reviewing campaign performance. A year-on-year traffic comparison against 2022 numbers is a comparison against a larger pool. Teams that reset their benchmarks to 2025 figures will get a more honest read of where they actually stand.
The revised 12.2% figure is the one to use in any planning document from here.
SOURCES & CITATIONS
FREQUENTLY ASKED QUESTIONS
What did SparkToro find about web traffic?
Was the original SparkToro figure different?
Does the data apply to Australian traffic specifically?
What does Fishkin say caused the decline?

James Dore is founder and editor of Prompt the Market and Strategy Director at Bushnote, a Sydney strategy, SEO and AI search agency. He has a background in strategy, media and financial services, and led consumer product marketing at MONEYME. He writes on AI search, answer engines, technology policy and media economics.







