Tracking breaks quietly. Here is the checklist that catches it before the quarter does.
Performance dropped three weeks ago and you still do not know whether it was the market, the creative, or a tracking break.
4 min read
Tracking almost never fails with an error message. It fails when a theme is updated, an app is removed, a checkout extension changes, or a consent banner is reconfigured. The first symptom is a performance decline that looks exactly like a media problem — so teams rebuild campaigns, refresh creative and change bids, chasing a fault that lives somewhere else entirely.
This audit takes about twenty minutes and needs no developer. Run it quarterly, and immediately after any theme change, app install or checkout change.
Step 1: place a real test order
Everything else is inference. This is observation.
Use a real product, a real checkout, and a discount code that takes it to a small amount rather than zero — a zero-value order sometimes behaves differently through checkout. Complete the purchase on a phone, not a laptop, because mobile is where most of your traffic and most of your tracking problems live.
Then watch what arrives, in this order:
- Meta Events Manager: does a purchase event appear within a few minutes, from both the browser and the server?
- Is exactly one purchase recorded, not two?
- Is the value correct, and in the ad account's currency?
- Google Ads conversions: does the purchase conversion register?
- Shopify analytics: is the session attributed to a source, or dumped into direct?
Step 2: check the volume ratios, not the absolute numbers
A single test order proves the path works. Ratios prove it works for everyone.
Pull the last 28 days and compare event counts to real order counts. You are not looking for a match — you are looking for a stable, sane relationship.
| Comparison | Healthy pattern | What a break looks like |
|---|---|---|
| Meta purchase events vs Shopify orders | Meta somewhat higher, stable week to week | A sudden step change in either direction |
| Browser vs server purchase events | Similar counts, deduplicated into one | Server events near zero, or double the total |
| Add to cart vs purchase | A consistent ratio for your store | Add to cart holding while purchase collapses — checkout tracking broke |
| Google Ads conversions vs orders | Lower than orders, stable | Sudden drop with no spend change |
The word doing the work is stable. Nobody can tell you what your ratios should be. Everybody can tell you that yours changed.
Step 3: match quality and identifiers
An event that arrives but cannot be matched to a person is only half useful. Check event match quality on your purchase event in Events Manager — below about 6 out of 10 means a meaningful share of your conversions are not being attributed at all, and your delivery system is learning from a thinner sample than it should be.
The fix is almost always the same: attach more customer fields to the event. Hashed email first, then phone, then name and address components.
On the Google side, the equivalent lever is enhanced conversions, which uses hashed first-party customer data to recover conversions that would otherwise go unattributed.
Step 4: consent, and why it is not optional
A consent banner that blocks tracking until acceptance will reduce your measured conversions. That is the banner working correctly. The failure mode to look for is the opposite: a banner that was reconfigured and now blocks everything permanently, or one that never actually applied the shopper's choice at all.
Check that accepted consent leads to events firing, and that declined consent leads to them not firing. Both directions. A setup that fires regardless of consent is a compliance exposure, not a measurement win.
Step 5: write down what you found
The audit's value compounds only if you keep the readings. A record of last quarter's ratios is what turns this quarter's numbers from raw data into a diagnosis.
Keep a single page with the date, the ratios from step two, the match quality scores, and any changes made to the store since the last audit. When performance moves unexpectedly, that page answers the first question — did tracking change? — in about thirty seconds, and lets you move on to the real cause.
Common questions
How often should I audit tracking?
Quarterly as a baseline, and immediately after any theme update, app change, checkout change or consent banner change. Those events cause most breaks.
How do I know if my Meta Pixel is working?
Place a real test order on a phone and watch Events Manager for the purchase event. If it arrives once, with the correct value and currency, within a few minutes, the path works.
My add-to-cart events are fine but purchases dropped. What broke?
Almost always something at checkout — a checkout extension, a theme update, or a post-purchase app. The events before checkout are unaffected, which is why the pattern is such a useful fingerprint.
Does a consent banner reduce my conversions?
Yes, and that is the banner doing its job. What you should check is that both outcomes work: accepted consent lets events fire, declined consent stops them.
Can this be monitored automatically?
Yes. Meta and Google both publish diagnostics about your existing setup, so event volume, deduplication and match quality can be watched continuously rather than checked once a quarter.
How Glimmio handles this
Glimmio watches Meta Pixel, Conversions API and Google Tag health continuously using the diagnostics those platforms already publish, so the quarterly audit becomes a monitored signal instead of a calendar reminder.
When a reading moves — event volume falling, deduplication stopping, match quality dropping — it surfaces as a recommendation with the underlying numbers attached, and nothing changes in your account unless you approve it.
- Manual approval by default — nothing runs unattended
- New campaigns and ads are always created paused
- 7-day recovery window on eligible changes
- 48 permissions across 13 roles, scoped per client account
Go deeper on this
The product pages and setup guides that cover what this article describes.
Searches this answers
- shopify tracking audit checklist
- how to check if meta pixel is working
- google tag not firing shopify
- conversion tracking broken ecommerce
- ecommerce tracking health check
Read next
Meta Pixel vs Conversions API: What Each One Actually Does
Your tracking setup was configured once, by someone who has left, and nobody can say whether it is still working.
ReadWhy Meta Says 340 Purchases and Shopify Says 214
Your ad platform and your store disagree about how many orders you got, and nobody can tell you which one is lying.
ReadThe Google Ads Search Terms Audit That Pays for Itself
Google Ads is spending steadily, conversions are flat, and you have no idea what people actually searched to get there.
Read