Get Your Tracking Together: Why AI Is Useless Without It

Freddy Kanter
May 8
3 min read

Let’s get something straight right off the line: AI is not magic. It’s not going to ride in on a hoverboard, point at your marketing spreadsheet, and say, “Ah yes, Sandra clicked the ad at 14:06 because she felt emotionally connected to your font choice.” No. What it can do — if you’re not asleep at the wheel — is take good data and turn it into something dangerously useful.

But here’s the catch: if your tracking is rubbish, your predictions will be too.

You wouldn’t try to drive a Ferrari with a blindfold on and one wheel missing, would you? Then why are so many businesses still relying on broken Google Analytics setups, pixel chaos, and form fills that go nowhere?

What Predictive Analytics Actually Means (No Buzzwords, Promise)

Predictive analytics is a fancy way of saying: “Can we use past behaviour to guess what people will do next — and sell them something before they know they want it?”

Sounds brilliant. But here’s the bit everyone forgets: AI doesn’t pull this data out of the clouds. It needs clean, complete, and consistent tracking. Garbage in, garbage out.

Want to know which channel brings in your most profitable customers?

Want to predict when someone’s going to drop out of your funnel?

Want to send the right offer at the right moment — not just some Tuesday in November?

Then your tracking needs to be tighter than a drum.

AI Can’t Guess What You Didn’t Measure

Here's the thing about machine learning: it’s incredibly smart at spotting patterns. But it’s also incredibly stupid if you feed it nonsense. If your analytics are riddled with gaps — events not firing, sessions misattributed, conversions tracked like it's 2006 — you’ll end up with AI models that look impressive but tell you absolutely nothing.

“Oh look, it says our top converting audience is ‘People Using Internet Explorer at 2am.’”

Well done. Let’s spend five grand on that.

The Power of Proper Tracking + AI

Now, when your tracking is in order — and I mean really sorted, with server-side tagging, accurate attribution, and no duplicate events — AI goes from being a gimmick to a growth engine.

You can:

Predict customer lifetime value before they’ve even converted
Forecast ROAS with eerie accuracy
Trigger smart campaigns that respond in real time to real behaviour
Build audiences that actually reflect intent, not just noise

Suddenly, you’re not just reacting. You’re anticipating. You're driving, not being dragged along behind your own data.

What You Should Do About It (Right Now)

Audit your current tracking setup — If you haven’t looked at it since Universal Analytics, start here.
Ditch vanity metrics — Nobody cares about impressions if nothing’s converting.
Get server-side tracking in place — Especially with cookie death looming.
Feed your AI models properly — With clean, structured, meaningful data.

And if all of that sounds like a faff — talk to someone who does it for a living. Like us. This is exactly the sort of thing we’re annoyingly good at.

Final Word

AI is changing predictive analytics faster than a Tesla in Ludicrous Mode — but it can’t do squat without the fuel of good tracking. So before you throw another buzzword into your boardroom, ask yourself: Is our data actually worth a damn?

Because if it’s not, the AI won’t fix it. It’ll just make the mess faster.