The Hidden Cost of Bad Tracking: How Google & Meta’s Modelled Conversions Drain Ecommerce Budgets

In 2025, data privacy is at an all-time high. Ad blockers, browser restrictions, and iOS updates have made traditional web tracking unreliable. In response, ad platforms like Google Ads and Meta Ads have leaned heavily on modelled conversions—using machine learning to “fill in the blanks” when actual conversion data is missing.

Sounds helpful, right? Until you realise:👉 You’re now paying for conversions based on estimates, not verified sales.

What Are Modelled Conversions?

Modelled conversions are estimated sales or leads that ad platforms assume happened, based on user behaviour patterns, traffic signals, and machine learning models. They exist because direct tracking (via cookies, pixels, etc.) is failing more often.

For Google Ads and Meta Ads, this means if your site’s tracking misses a sale, they’ll still attribute it based on what they “think” happened.

Example:

• User clicks your ad

• Tracking breaks (e.g., blocked by iOS, browser, or bad setup)

• Google/Meta “model” that a conversion likely happened

• You pay for that conversion, even though it wasn’t actually tracked

Why Bad Tracking Amplifies the Problem

If your ecommerce site relies only on browser-based tracking (pixels, tags, scripts) without server-side tracking (S2S), you’re giving Google & Meta no choice but to model even more conversions.

The more gaps in your data, the more guesswork they’ll do.

For advertisers, this leads to:

• Over-reported conversions

• Higher CPA (Cost Per Acquisition)

• False sense of ROAS (Return on Ad Spend)

• Wasted budget on assumed sales

How Server-Side Tracking Fixes This

Server-side tracking captures conversion events directly from your server, bypassing many browser limitations. This creates a clean, verified stream of conversion data back to Google & Meta.

The result:

• Fewer modelled conversions

• More accurate attribution

• Better optimisation decisions

• Lower wasted spend

Real-World Impact: What Brands Are Seeing

Brands that implement server-side tracking with unified attribution platforms like Shoplytics consistently see:

• Cleaner ROAS reporting

• Lower CPA by reducing ghost conversions

• More reliable data for smart bidding algorithms

• Confidence in scaling ad spend without data leaks

Are You Paying for “Phantom” Conversions?

If you haven’t upgraded your tracking setup, chances are you’re already paying for modelled conversions you could avoid.

The good news? Fixing it isn’t as hard as it sounds.

👉 Book a free, no-obligation tracking health check with Shoplytics. We'll show you exactly where your tracking leaks are—and how much you might be losing.

No sales pitch. Just data clarity.