The problem with broken tracking is that nothing looks broken
When tracking breaks, Google Analytics does not show an error. The reports still load. The charts still have lines. One of the numbers is just wrong, and nobody knows which one.
So the team keeps making decisions on it. Budgets move. Campaigns get paused. A checkout change gets blamed for a drop that never happened.
An audit is how you find this before it costs you. This post covers what to check, how to check it by hand, and how to get the same answer in about 30 seconds.
What a Google Analytics audit actually checks
An audit answers one question: is what Google Analytics recorded the same as what happened? Six checks cover most of it.
- Is every step of the journey being recorded? Product view, add to cart, checkout start, payment, purchase. A step that records nothing leaves a hole in every funnel report.
- Do the counts make sense against each other? If 21 people bought, at least 21 people reached the payment step. When a later step is bigger than an earlier one, something upstream is not firing.
- Are the numbers in their usual range? A step that normally records about 3,900 events a week and now shows 3,400 is either a real drop or a tracking fault. You need to know which.
- Are conversions reaching your ad platforms? Google Ads optimises on the conversions it receives. If the checkout conversion stops arriving, bidding keeps running on old signals.
- Does revenue match your store or CRM? Compare purchase count and revenue for the same dates. A small gap is normal. A large one is a tracking problem.
- When did it start? The date matters as much as the fault. It tells you which reports to stop trusting, and usually which release caused it.
How to audit Google Analytics by hand
You can do all of this yourself. Set aside two to three hours for a first pass.
- List the events you expect. Write down every step a customer takes and the event name that should record it. If nobody can produce this list, that is your first finding.
- Check each event exists. In Reports, open Engagement, then Events. Set the date range to the last 7 days. Every event on your list should be there with a count above zero.
- Compare against the week before. Add a comparison period. Look for events that dropped to zero, or moved far more than your traffic did.
- Walk the funnel yourself. Open DebugView, go through your own site from product page to payment, and watch whether each event fires as you do it.
- Check the key events. In Admin, open Key events and confirm the ones your business cares about are marked.
- Check the Google Ads link. In Google Ads, open Conversions and look at the status and the last date each conversion was recorded.
- Match purchases to your store. Pull order count and revenue from Shopify or your CRM for the same 7 days and put them next to the Google Analytics numbers.
It works. The trouble is that it is a one-off. Tracking breaks on a release day, and nobody repeats a three-hour check every morning.
The 30-second way
The video at the top shows the same audit in ClickBoss. You type one line:
Audit my Google Analytics
ClickBoss compares each step of the journey against what it expects from your own recent data, and answers in plain language. This is the result from the real account in the video:
3 tracking problems; 2 affect money. The biggest: Payment step has recorded nothing since 21 Aug 2026, while 21 people bought.
Underneath is the checklist, problems first:
| Check | What it found | Impact |
|---|---|---|
| Payment step | Not recorded. 0 events against about 9 expected, while 21 people bought | 21 buyers affected |
| Google Ads checkout conversion | Not recorded. 0 events against about 26 expected | 21 buyers affected |
| Product views | Below usual. 3,376 events against about 3,871 expected | About 495 events |
| Add to cart | Recording as usual | None |
| Checkout starts | Recording as usual | None |
| Purchases | Recording as usual | None |
Two things in that answer are worth noticing.
It gives you the date. "Nothing since 21 Aug" tells you exactly which weeks of funnel reports to stop trusting.
It says when it does not know. On product views the result reads: "either a real drop or tracking, the data can't tell which." That is the honest answer, and it tells you where to look next.
What to do with the results
- Fix what affects money first. A missing payment step or a missing ads conversion changes bidding and budget decisions today. A dip in product views can wait a day.
- Hand it over as it is. Switch to the Table view and export to CSV. Your developer or agency gets the exact events, the counts and the dates, with nothing to interpret.
- Mark the bad period. Note the start date against any report that used the broken step, so nobody quotes those numbers later.
Keep it watched
An audit is a photo. Tracking breaks again the next time someone ships a release.
In ClickBoss, save the audit and switch on Watch daily. It runs the same check every day and alerts you when a number moves beyond normal. You find out the morning after it breaks, not at the end of the quarter.
FAQ
How often should I audit Google Analytics?
After every site or app release, and on a schedule in between. Releases are when tracking breaks most often. Daily is realistic if the audit is automated. By hand, monthly is the minimum, because a fault found after a quarter has already shaped a quarter of decisions.
What is the most common thing an audit finds?
A step that silently stopped recording after a change to the site, most often in checkout or on a form. The page works and people still buy, but the event that records it no longer fires, so every funnel report from that day on shows a drop that never happened.
Does ClickBoss change anything in my Google Analytics?
No. ClickBoss connects to Google Analytics with read-only access. It reads your events and reports what it finds, and it cannot change your settings, events, key events or data. Fixes stay with you, your developer or your agency, and the audit tells them which step to fix and since when.
Can I share the results with my developer or agency?
Yes. Open the Table view under the answer and export it to CSV. The file lists each check, what it found and its impact, with the date range on the first line, so your developer or agency can start fixing without a call to explain the numbers.