5 Signs Your Tracking Is Broken (Without You Noticing)

The worst kind of ad problem is the invisible one. Broken tracking makes you optimize on wrong data, so you pause a good ad and scale a bad one — convinced you’re doing the right thing. Here are five signs your tracking setup needs an urgent review.

1) Platform numbers differ sharply from store sales

If Meta says it drove 100 orders and your checkout says 60, you have a gap. A small gap is normal due to attribution windows, but a large, persistent gap means events are recorded wrong — either missing or duplicated. Track this ratio weekly; if it swings wildly, the tracking isn’t reliable.

2) Purchase events fire more than once

If the Pixel fires a “Purchase” event twice (once from the browser and once from the server without proper deduplication), you’ll see phantom sales. This makes ROAS look better than reality and misleads the platform’s optimization. A common tell: the purchase value on the platform is roughly double the real value.

3) The thank-you page doesn’t always record the conversion

In many stores (especially with cash on delivery or a redirect to an external gateway), the confirmation page doesn’t always load, so the event is lost. If recorded conversions are far below actual orders, you’re losing important signals the platform learns from.

4) GA4 tells a completely different story

If GA4 matches neither the platforms nor the store, there’s likely a misconfiguration: duplicate tags, untuned cross-domain settings, or missing e-commerce events. GA4 should be the neutral referee; if it’s unreliable too, you have no source of truth.

5) Performance “dropped” suddenly after privacy updates

If you noticed a sudden drop in recorded conversions after an iOS update or after enabling a consent banner, you’re most likely losing browser-side signals. This is exactly what CAPI (server-side tracking) exists for — it recovers a large share of the lost signal.

How to check quickly

• Compare checkout orders to platform numbers for a week.
• Use Pixel inspection tools (like official browser extensions) and look for duplicate events.
• Confirm CAPI is live and that deduplication uses a shared event_id between browser and server.
• Review GA4’s e-commerce setup and confirm purchase events arrive with correct values.

A sixth sign: purchase value isn’t recorded correctly

In many cases the Pixel fires a purchase event but without a value, or with a wrong fixed value (like 1 or 0). The result is that the platform can’t tell expensive orders from cheap ones, so it optimizes for “count” instead of “value.” If the average conversion value on the platform looks unreasonable compared to your real average order, that’s a sign value isn’t being passed correctly — and it silently wrecks optimization.

A 30-minute audit plan

You don’t need a big project to catch most of these problems. Do the following in order:

• 10 minutes: place a test order on your store, open a Pixel inspection tool, and confirm the purchase event fires once with the correct value and currency.
• 10 minutes: compare checkout orders over the last 7 days to the conversions reported by Meta and Google, and record the gap as a percentage.
• 10 minutes: open GA4 and check that e-commerce events (view_item, add_to_cart, purchase) exist and arrive with a value. If they’re missing, you have a setup problem.

If a large gap or duplicate events show up, don’t touch budgets before fixing tracking — any optimization on top of wrong data magnifies the error instead of solving it.

And remember tracking isn’t set-and-forget; every site, theme, or new-tool update can break an event. Make a tracking check part of your monthly routine, not an emergency after performance drops.

What usually breaks tracking?

Most tracking problems aren’t bad luck — they have recurring causes you can avoid once you know them:

• More than one Pixel on the same page: apps or themes install the Pixel twice, so events duplicate and inflate.
• Theme updates or thank-you page edits: any change to checkout or the confirmation page can remove the code without you noticing.
• External payment gateway: if the customer leaves to another gateway and returns, the event can be lost or recorded on the wrong domain.
• A misconfigured consent banner: it can block events even from people who actually consented.
• Incomplete CAPI setup: server-side tracking without a shared event_id creates duplication instead of fixing it.

The good news is that most of these are caught with a simple check and fixed once, correctly. The bad news is they recur with every update, so if your store evolves constantly, keep a documented reference for your tracking setup and review it after any major change. It saves budget that would otherwise burn on decisions built on wrong data, and lets the platform’s algorithm learn from clean signals.

Bottom line

Tracking isn’t a technical detail — it’s the foundation every ad decision is built on. Clean data = right decisions = budget spent where it belongs. If you suspect one of these signs applies to you, Madar’s free consultation can diagnose your tracking stack and tell you exactly where it breaks.