How to check if traffic is real: three places to look, each catches what the others miss.
1. Analytics: what people do
Bot and fake traffic leave these marks (Cloudflare):
- abnormally high pageviews
- abnormally high bounce rate
- surprisingly high or low session duration
- junk conversions (gibberish emails, fake names)
- a spike from an unexpected location
In Google Analytics: how to identify bot traffic in GA4. GA4 already drops known bots, so what you see is the part it did not recognize.
2. Server logs: who is asking
Analytics sees only visitors that run its script. Logs see every request. For the visits you doubt, check:
| field | real visitor | suspect |
|---|---|---|
| user agent | a current browser | empty, a library name (python-requests, curl), or an old browser version |
| IP address | home, mobile or office network | one range hitting thousands of times, a datacenter |
| requests | a few pages, seconds apart | the same page, or pages in machine-speed order |
| claimed crawler | verifies | says “Googlebot” but fails the check below |
Verify a claimed Googlebot: Google says its common crawlers use published IP ranges and the reverse DNS name matches crawl-***-***-***-***.googlebot.com or geo-crawl-***-***-***-***.geo.googlebot.com (Google). AI bots publish ranges too: AI crawlers. Paste log lines into the bot traffic checker to test them all at once.
3. Ad reports: what you paid for
If you buy traffic, read the invalid traffic numbers. In Google Ads, add the Invalid clicks column; Google removes invalid traffic found before billing and credits what it finds later (Google Ads Help). That covers what Google catches. Clicks by real people paid to click, or click farm traffic, can pass as normal. Background: invalid traffic.
What to do with the result
| finding | action |
|---|---|
| a bot you do not want | block it at the firewall or CDN by user agent, IP range or behavior |
| an AI crawler | block it or allow it on purpose |
| invalid ad clicks | report them to the ad platform, exclude the placements |
| a protection that blocks real research bots | check what protects the site |