Open the analytics and the number is up. Traffic doubled. Somebody says the new site is working. Then the phone stays as quiet as before.
Before anyone celebrates a traffic number, it is worth knowing how much of it was ever a person.
Key takeaways
- On a new service business site, most recorded visits are machines, not customers
- Google Analytics counts browsers, not people. One owner on three devices is three users
- Two columns tell the truth: engaged sessions and the city the visit came from
Where the traffic actually comes from
We pulled the numbers on a site we run. Ten days, 74 visits. Sorted by city, the top entries were Forest City, Dublin, Luleå, Singapore and Prineville. Every one of those is a town with a Meta data center in it. Every one of those visits lasted zero seconds.
That is not fraud. When an ad is published or edited, Meta loads the landing page from its own servers to review it, then again for the link preview, then again to prefetch it for people it expects to tap. None of those loads is a customer. All of them count as a visit.
Strip those out, plus the one second visits from random cities that scanners produce, and 74 visits became four people who could plausibly have been customers. The dashboard said traffic was up 47 percent.
A traffic number without a filter is a count of how many machines looked at the page.
Why the user count is not a count of people
Google Analytics does not know who anyone is. A user is a cookie in one browser. The same owner on a laptop, then a phone, then a private window, is three users. Clear the cookies and a fourth appears.
On a small site, the owner and whoever is building it load the page more than anyone. Most exclusion methods only cover the one browser they were set up in. The phone still counts.
The two columns that tell the truth
Google only counts a session as engaged if it lasted ten seconds, viewed two pages, or completed a goal. A crawler does none of those. And a visitor from a town you have never advertised in, who stayed zero seconds, did not come to buy a roof.
| What the report shows | What it usually is |
|---|---|
| Dozens of visits, average time on site under five seconds | Ad platform review crawlers and link scanners |
| One visit each from twenty cities on four continents | Automated scanners hitting the home page once |
| A cluster of long, engaged visits from your own city | You, your team, and whoever built the site |
| A handful of visits from towns in your service area, ten seconds or longer | The people who might actually call |
Read your own report in ten minutes
Read engaged sessions, not users Reports, then Acquisition, then Traffic acquisition. Ignore the users column. Engaged sessions and average engagement time are the closest thing to a count of people.
Sort by city Reports, then User attributes, then Demographic details, dimension set to City. Zero seconds from a place you do not serve is a machine.
Mark your own devices Exclude your traffic on every device you use, not just the office computer. Then load the site from your phone and confirm it does not appear in the realtime report.
Ask the only question that matters Of the real people in your service area, how many called, booked, or filled out the form. That is the number to put next to the ad spend.
Why is my website traffic coming from other countries?
Most of it is automated. Ad platforms, search engines, security scanners and link preview tools load pages from data centers around the world. If the engagement time is zero, they were not people.
Does bot traffic hurt my rankings or my ads?
Not directly. The harm is in the decisions. If the report says traffic doubled and the budget goes up because of it, the machines just spent the money.
From a visit count to a booked revenue count
Analytics will report visits and ad platforms will report clicks, and both can grow every month while the schedule stays flat. Booked revenue is harder to count, and it is the only number the business runs on.
Getting there means wiring the site, the phone, the booking calendar and the job records together so a visit that became a job can be traced back to the ad that produced it. That is what Metric Minds Marketing calls the Revenue Circuit. You cannot trace a job back to a visit that was never a person.
Sorting the machines from the people is one of the first things Metric Minds Marketing does in any Revenue Forensic Audit.
Sources: Google Analytics Help, Engaged sessions · Meta for Developers, Meta web crawlers