Your Google Analytics Is Probably Counting the Wrong Things

Google Analytics is free, it installs in minutes, and it produces a dashboard full of numbers that look authoritative. That last part is the problem. An account can be set up, left alone for a year, and quietly report things that are not true: traffic that is not people, popularity that is really the owner checking the site, and “conversions” that are not leads.

None of these problems is exotic. They are easy to create and easy to miss, and every one of them makes the numbers look better or worse than reality in ways that lead to the wrong decisions. Here are the three we look for first when we open an account, how to spot each, and how to fix them.

Problem one: robots that look like visitors

Google Analytics 4 filters out known bots, but only the ones on published lists. Automated browsers that run a page’s JavaScript, the way a real browser does, get counted as people.

They are easier to spot than you might think, because they behave nothing like an audience. A typical pattern: a country with no obvious connection to your business suddenly sends a large share of your traffic. Every one of those “users” visits exactly once, stays a fraction of a second, arrives with no referring site, and somehow visits every page you have, including the cookie policy. Real people do not read cookie policies in bulk.

A robot like this does no harm to your site. The harm is to your decisions. If it makes up a third of your sessions, a month where real visitors fell can look flat, and a campaign that did nothing can look like it worked.

How to check yours

Open Reports, then the demographic report by country, and add two columns if they are not there: Engaged sessions and Average engagement time. You are looking for this signature:

• Sessions and users almost equal: every “person” visits exactly once.
• Engagement time close to zero.
• Almost no engaged sessions.
• A country with no plausible reason to be interested in a Cayman business.

The opposite signature is also automated: one “user” with dozens of sessions, each lasting twenty minutes or more.

Analytics will not let you delete this traffic after the fact. What you can do is compare with and without it: use Add comparison to exclude the country, and read that version of every report from now on.

Problem two: you are your own biggest visitor

Every small business does this. The owner checks the site every morning, staff open it to copy links for customers, a web designer reloads the same page forty times while fixing it. On a site with tens of thousands of visitors a month that is noise. On a Cayman business site with a few hundred, it can be a sizeable share of the total, and it lands on exactly the pages you care most about: the home page, the About page, the page you have just changed.

The giveaway is views per user. If a page shows seventy views from fifteen people, that is not seventy readers. It is a handful of readers and somebody on your team.

How to fix it

Analytics has a built-in route: define your office’s internet address as internal traffic in the web data stream’s tag settings, under Admin, then switch the Internal Traffic data filter from testing to active. It works well if your team sits in one office with a fixed connection.

It works poorly if your team is spread out, works from phones, or has an address that changes, which describes most small teams. The alternative is a small change to the website itself, so that browsers belonging to your team do not load the analytics tag at all: anyone logged in to the site’s admin is excluded automatically, and a private link excludes a phone or a second browser. It is a short job for whoever built your site, and it is how we handle it for ourselves.

Problem three: counting something that is not a lead

Analytics lets you mark certain events as key events, which used to be called conversions. They are the numbers that are supposed to tell you whether the website is doing its job.

A common mistake is marking form_start. That event fires when someone clicks into a form field. It does not mean they filled anything in, and certainly not that they sent it. An account can report twenty of these a month, and to anyone reading the dashboard quickly they look like twenty leads.

The opposite mistake is just as common: the thing that actually starts most conversations with a Cayman business, someone tapping the WhatsApp button or the phone number, is not counted at all. Often the website is sending the right signal and the tag manager that forwards events to Analytics simply has no instruction for it, so it goes nowhere, for months, and nobody notices because nothing is visibly broken.

How to check yours

1. Decide what a lead is for your business. For most in Cayman it is some mix of a WhatsApp message, a phone call, a form actually submitted and a booking made.
2. In Admin → Events, look at what is marked as a key event. If it is form_start, page_view or scroll, it is not measuring leads.
3. Open the Realtime report on your computer, then tap your own WhatsApp button and phone number on your phone. If no event with a sensible name appears within a minute or so, those taps are not being recorded.
4. Once the right events arrive, mark them as key events and unmark the ones that are not leads.

Step three is the one that matters. It takes two minutes and it is the only way to know for sure.

What changes once the numbers are clean

Clean numbers are usually smaller, and that is the point. Once the robots and your own visits are out, three things become visible that were hidden before:

• The real trend. A month-on-month change you can believe, even when it is not the change you wanted.
• Your local audience. Filter to the Cayman Islands and look at engaged sessions. For most businesses here it is the number that pays the bills, and it is often a fraction of the headline total.
• Which pages produce enquiries. Not which pages get views, but which ones people were on when they tapped WhatsApp or sent the form. That is where to spend your time.

Questions we get asked

Does bot traffic hurt my Google ranking?

There is no evidence that robots visiting your site affect how Google ranks it. The damage is to your decisions: you think a campaign worked, or a page is popular, when neither is true.

Should I switch to a different analytics tool?

Not for this. Every analytics tool has to decide what counts as a person, and all of them can be fooled by browsers that behave like people. The fixes above are about reading the tool you have correctly.

What single number should I watch?

Engaged sessions from the countries you actually sell to, plus the count of genuine key events. Total visitors is the number most distorted by robots and by your own team, which makes it the least useful headline you can have.

My developer set up Analytics years ago. Is it still working?

Possibly not the way you think. Tags break when sites are redesigned, when cookie banners are added, and when someone edits the tag manager without testing. The two-minute Realtime test above is the quickest way to find out.

Where to start this week

Do the Realtime test: open the Realtime report, tap your own WhatsApp and phone links from your phone, and see what arrives. Then look at your countries with the engagement columns switched on. Those two checks will tell you whether your numbers mean what you think they mean.

If you would rather have someone go through the account with you, measurement is part of what we do: digital marketing for Cayman Islands businesses, set up so the numbers can be trusted. And our free checker at toctoc.ky/seo-checker will show you how your site itself looks to Google and AI assistants.