Most ad accounts I audit do not have an ads problem. They have a tracking problem.
The campaigns look fine. The budget is being spent. But the numbers in Google Ads, Meta and GA4 do not match what is really happening in the business. And when the data is wrong, every decision built on it is wrong too.
Here are the three mistakes I find again and again, and how to fix each one.
Mistake 1: Counting the wrong thing as a conversion
This is the most common one. The account says it got 120 conversions last month. The sales team says they got 40 leads. Both are telling the truth.
What usually happens:
- A page view of the thank-you page is counted, so every refresh counts again.
- The same lead is counted twice, once by the Google tag and once by an imported GA4 event.
- Small actions like clicking a phone number or starting a form are set as primary goals.
When this happens, smart bidding learns from fake wins. It finds more people who click, not more people who buy.
The fix:
- Pick one primary conversion for each real business goal, like a submitted lead form or a purchase.
- Mark everything else (clicks, scrolls, form starts) as secondary, so you can watch it without bidding on it.
- Fire the conversion on the real success event, not on a page view, and use a unique ID (like an order or lead ID) to stop double counts.
- Open GA4 DebugView and Google Tag Assistant, submit a test lead, and check it fires exactly once.
Mistake 2: Relying only on browser tracking
A few years ago, a pixel in the browser was enough. Not anymore.
Ad blockers, privacy settings on phones, cookie limits in browsers and cookie banners all stop browser tags from sending data. So a real share of your conversions never reach Google or Meta. The platforms see fewer results than you actually got, and they optimise with half the picture.
The fix: send important events from your server as well as the browser.
- Meta Conversions API (CAPI) sends leads and purchases from your server straight to Meta.
- Google enhanced conversions add hashed customer details (like email) so Google can match more conversions.
- Server-side Google Tag Manager gives you one clean place to manage all of this.
- Consent Mode v2 keeps you respectful of user choices while still giving the platforms modelled data.
Done right, this does not mean tracking people more. It means the data you are allowed to collect actually arrives.
Mistake 3: No link between a lead and a sale
This one costs the most money, and almost nobody notices it.
If your ads only learn from form fills, they will happily bring you cheap leads who never pick up the phone. A ₹200 lead that never buys is more expensive than a ₹900 lead that becomes a customer.
The fix: close the loop between your ads and your CRM.
- Save the click IDs (GCLID for Google, fbclid for Meta) and UTM tags in hidden form fields.
- Store them with each lead in your CRM or even a Google Sheet.
- When a lead becomes qualified or turns into a sale, send that event back to Google (offline conversion import) and Meta (CAPI).
- Add a value to each stage, so the platforms learn which leads are worth the most.
Now your ads stop chasing clicks and start chasing customers.
A quick 5-minute self-check
- Does your number of conversions roughly match your real leads or sales?
- Is there only one primary conversion per goal?
- Do you send key events from the server, not only the browser?
- Can you tell which campaign each paying customer came from?
If you answered no to any of these, your ads are working with blurry data. The good news: every one of these can be fixed.
Want me to check your setup? I review ad accounts and tracking for businesses that want sales, not just clicks. Book a free call or message me on WhatsApp.
— Samrat