What do you measure when AI bots inflate your devtool traffic?
My team was pretty happy in December.
Direct traffic jumped 150% for one of our cloud devtools.
Then we actually looked at the data.
I wrote about developer marketing metrics back in Nov 2025. A month later, the conversation changed. Suddenly everyone was talking about AI bots showing up as traffic, often from China and Singapore. Docs pages getting hammered. Sessions going up. Very little downstream activity.
So now I’m genuinely unsure what the safe bet is for traffic metrics in devtools.
Two options we keep circling back to.
One is filtering. Excluding China and Singapore from analytics gives cleaner charts, but it’s a weak assumption. AI traffic is not tied to geography anymore, and filtering by country feels like sweeping the problem under the rug.
The second is shifting attention away from traffic altogether. First API calls. Successful auths. SDK installs. Repeated usage. Things that require a human to actually do something.
Right now, we’re leaning toward treating traffic as a directional signal only, and using product actions as the real indicator of developer interest.
Curious how others here are handling this. What metric do you trust today when traffic numbers start lying?
Replies