Huge increase in Windows traffic

[ace (Adam Engst)]

With the discussion about the iOS 26 adoption rate, I decided to check Google Analytics to see what it says. Who knows about iOS, because I couldn’t get past the fact that Windows is now making up the bulk of our traffic. How?!? Why?!?

The first screenshot is for TidBITS; the second for TidBITS Talk. Windows is the blue line in the first one and green in the second one.

This Windows traffic just started in September 2025. I had a conversation about it with the Analytics Advisor chatbot since I find Google Analytics almost entirely incomprehensible. It thinks the traffic is bot traffic, though presumably some new bot that Google isn’t automatically filtering out. Overall, it was a good experience, since it would have taken me a long time to think up these comparisons and figure out how to find them.

I have conducted a broader analysis of the Windows traffic increase since September, and I’ve found a significant commonality. While I could not isolate a specific channel or landing page that is responsible for the increase, I discovered a dramatic shift in user engagement that points to a clear explanation.

Since September, the engagement of Windows users has changed dramatically:

• The bounce rate has increased from 57.7% to 91.8% . This means that the vast majority of new visitors are leaving your site after viewing only a single page.
• The average session duration has plummeted from 107.5 seconds to just 17 seconds . This indicates that these new visitors are not engaging with your content.

Bounce rate, Average session duration for Windows

Sep 1, 2025–Jan 8, 2026 and May 1–Aug 31, 2025

Bounce rate

91.8%

+59.0%

Average session duration

16s

-84.2%

This pattern of high bounce rates and low session duration is a strong indicator of bot traffic . It appears that automated bots, not human users, are the source of this increase in Windows traffic. This would explain why the traffic is coming from various sources and landing on different pages without a clear pattern, and why our previous analysis of traffic from Singapore showed similar characteristics.

While I cannot definitively confirm the source of the bot traffic, the data strongly suggests that the increase in Windows traffic is not from genuine users interested in your content.

[tommy (Tommy Weir)]

So my low figures on my sites are actually even lower… sigh.

The analysis makes sense. Google should be able to filter that into a separate bot line on the chart.

[Dan_W]

@ace , when Google tells you that it has detected traffic from Windows clients, it’s basing that on the easily spoofable browser ID string that is passed in by the client during requests.

The majority of traffic (a much as 70-80%, in my experience) to any web site nowadays is generated by crawlers looking for content with which to teach an LLM. Google cannot automatically detect when this is the case, simply because it’s too easy to spoof in undetectable ways. Many sites require Google’s tracking cookies to be set, so the crawlers are smart enough to do this and other things in order to get to the content. It’s very likely that the crawlers are just using Windows in their browser ID because it’s generic; they don’t want to use a different value which might affect the content they get back.

[Simon]

Any chance this bot onslaught is related to the sluggishness of the website as of late? Page loads and page updates after edits appear to be taking significantly longer here than they used to.