Every marketing decision you make leans on your analytics. Which channel works, which page converts, whether the ads are paying for themselves. So it’s worth asking a question most businesses never do: is it actually recording what happens?
On the sites we check, the honest answer is often “partly”. Here’s what we find, anonymised, and how to check your own.
1. A tag that records nothing
One site we look after had a Google Analytics tag added to every page six months earlier. When we checked, Google itself answered that tag’s address with “not found”. The property behind it had gone, or the ID had been wrong from the start. For six months it had loaded on every visit and recorded precisely nothing.
Nobody noticed, because the site had other tags that were working, so the reports never looked empty. Somewhere, someone may have been waiting for figures from a property that would never have any.
2. Three people, three ways of adding tracking
It’s common to find tracking added through a WordPress plugin, again through Google Tag Manager, and once more by pasting code into the page, usually by different people at different times. Sometimes that double-counts visits. Sometimes, as on the site above, it doesn’t, and only looks like duplication.
That second case is the trap. We nearly fell into it ourselves: removing what looked like a duplicate would have switched off the only copy feeding two of the site’s reporting properties. We measured before and after, spotted it within minutes and put it back. The lesson: never remove tracking because it looks duplicated. Measure what each piece actually sends.
3. Visits your cookie banner hides
Under UK rules, most analytics shouldn’t record a visitor until they’ve agreed to cookies. That’s correct. It also means that when we test sites, a visitor who ignores the banner often sends no analytics data at all.
Your reports therefore show the people who clicked “accept”, not everyone who came. That’s fine as long as you know it. It’s a problem when you compare this year’s figures with a year when there was no banner, or when a banner change quietly moves the numbers and nobody connects the two.
4. Ads that can’t prove what they sold
A redirect that drops the tag on the end of an ad link, a form whose thank-you page isn’t counted, a conversion counted twice: each makes your ads look better or worse than they are, and your bidding follows the wrong signal. We wrote about the most common ones in Why your Google Ads aren’t working.
5. Nobody knows which property is the real one
Old agencies, old logins, a property set up “just to test”. It’s common to find two or three analytics properties receiving data from one site, with the business owner only able to see one of them. The figures a previous agency reported may have come from a property you can’t even open.
How to check yours in fifteen minutes
- Open your analytics’ real-time view, accept the cookie banner on your own site on your phone, and browse three pages. Do you see yourself? Once, or twice?
- List every property and tag your site loads. Google’s free Tag Assistant shows them. Can you log in to each one? Do you know who set it up?
- Submit your own contact form. Does it show up as a conversion, once?
- Click one of your own ads and look at the address you land on. Is the tag on the end still there?
- Write down when your cookie banner went live, and remember it every time you compare this year with last.
If any of that turns up something odd, that’s the kind of audit our digital marketing and paid media work starts with, because every decision after it depends on the numbers being right. Tell us what you’re seeing.
Got a problem like this? Good. We like those.
Tell us what you’re dealing with and we’ll tell you honestly what we’d do.
