The most expensive problem in most ad accounts is not the ads. It is the measurement underneath them.
Over the past few weeks I have looked inside a lot of startup ad accounts. Different industries, different budgets, different platforms. But one problem shows up more often than any other, and it is never the one the founder thinks they have.
It is not the ad creative. It is not the targeting. It is not even the budget.
It is conversion tracking that quietly broke months ago, and nobody noticed.
Here is the pattern I keep finding. The Google Ads tag is installed and firing, so everything looks healthy in the dashboard. But it fires when the page loads, not when someone actually books a demo or signs up. The form is often an embedded tool like HubSpot or Typeform that submits without reloading the page, and no event was ever wired to that submission.
The result: the ad platform counts visitors as conversions, or counts nothing at all. Either way, the number your dashboard calls "conversions" is not measuring what you think it is measuring.
This is not a rare edge case. It happens in ordinary, predictable ways. A website gets rebuilt and the tags do not come across. Someone switches form providers and the old conversion event points at a page that no longer exists. A tag manager gets tidied up and a trigger is deleted. Cookie consent gets added and quietly blocks the tags for a chunk of visitors. Each change is small. None of them announces itself. The dashboard keeps showing numbers, and numbers feel like truth.
Modern ad platforms are optimisation machines. Google and LinkedIn take whatever conversion signal you give them and steer your budget toward more of it. That is their entire job.
So if your conversion event fires on page load, the platform learns to find people who load pages. It gets very good at this. You get more clicks, more visits, cheaper "conversions" in the dashboard, and the same number of actual demos as before. The machine is working perfectly. It is just optimising toward the wrong target, because that is the target you gave it.
Every optimisation decision made on top of that data inherits the problem. Pausing the "underperforming" campaign that was actually producing your best leads. Scaling the one that produces cheap clicks and nothing else. Testing landing pages against a metric that does not connect to revenue. Months of work, steering by a broken compass.
Here is where it gets genuinely costly. After a few months of spend with flat results, the founder draws the obvious conclusion: paid does not work for us. The budget gets cut, the channel gets written off, and the company goes back to relying on channels it has already maxed out.
But paid was never given a fair test. The measurement was lying from day one. Writing off a channel based on broken data is like concluding a car does not work because the fuel gauge was stuck on empty.
I would go as far as saying this: for early-stage companies, a broken tracking setup is more expensive than bad ads. Bad ads waste this month's budget. Broken tracking corrupts every decision you make for as long as it stays broken, and then poisons the conclusion you draw at the end.
You do not need an expert for the first pass. Three checks will tell you most of what you need to know.
1. Fire a test conversion yourself. Open your site in a private browser window, click through your own ad if you can, and complete the form or signup exactly like a customer would. Then check whether that conversion shows up where it should, in Google Ads or LinkedIn Campaign Manager, attributed to the right campaign. If your own test does not show up, neither do your customers.
2. Compare two numbers that should roughly match. Take conversions reported by your ad platform for last month, and compare them with the demos or signups your CRM or calendar actually recorded from paid traffic. They will never match perfectly. But if the platform claims 80 and your calendar shows 12, you have found your problem.
3. Check what the conversion event actually is. In Google Ads, look at what your primary conversion action is tied to. If it says something like "page view" or points at a thank-you page that no longer exists, the platform has been optimising toward noise.
If all three checks come back clean, genuinely good news: your foundation is solid and your optimisation decisions are standing on real ground. You are ahead of most.
Fix the measurement before you touch anything else. Not the creative, not the targeting, not the budget. Any change you make while the data is broken just adds another layer of noise to untangle later.
The repair itself is usually not a big project. Wiring conversion events to real actions, cleaning up the tag setup, and validating the data end to end is days of focused work, not months. The hard part is knowing exactly what is broken, in what order to fix it, and how to prove it is fixed. That is precisely the kind of thing a structured audit surfaces in the first couple of days.
If you suspect your numbers would not survive the three checks above, that suspicion is worth acting on. It is the cheapest problem in your ad account to fix, and the most expensive one to ignore.
If you want a second pair of eyes on it, that is exactly what I do. Get in touch and I will show you where to look first.
Book a free 30-minute strategy call. We'll look at your funnel together and I'll give you my honest read on the biggest opportunity, whether or not we work together.
Choose which cookies you allow. Essential cookies are always on because the site needs them to work. You can change this anytime.