GA4 purchases suddenly dropped — put them next to your order count before you fix anything
When GA4 purchases or revenue fall off a cliff one week, GA4 alone cannot tell you whether you lost customers or lost tracking. Put your store admin order count next to it and divide GA4 purchases by orders, week by week. A broken week shows up as a step down. Here is how to read it and what to check in that week.
- Reading the numbers
- GA4
In an earlier post (Why your order count and conversion count never match), we explained that the order count in your store admin and the purchase count in your analytics tool always differ by a few percent up to around a tenth. They come from different layers, and neither is broken. This post is about what to do when that gap suddenly changes.
You open GA4 and purchases or revenue have dropped sharply from one week on. Did people stop buying, or did GA4 stop counting? Looking at GA4 alone, you cannot tell. Both show up in exactly the same way: fewer purchases.
What separates them is a single ratio. Here is how to use it.
1. Look at the order count in your store admin first
GA4 purchases belong to the visit layer. Orders in your store admin belong to the order layer. The order layer does not depend on browsers or tracking tags. If an order comes in, it is recorded as an order.
So for the week GA4 purchases fell, check the order count for the same period in your admin.
- Orders fell by the same amount — you lost customers, or they stopped buying. Tracking is working.
- Orders did not fall — what dropped is not sales but the counting. Suspect tracking.
Check: line up the same dates and the same time zone. If your GA4 property and your store admin use different time zones, the day boundaries shift and the weekly totals drift slightly.
2. Track GA4 purchases ÷ orders, week by week
The order count alone will not settle it when the drop is modest. So divide GA4 purchases by admin orders and lay the ratio out week by week.
This ratio is never exactly one. Some visitors block tracking, some orders are entered by hand or taken by phone, and some shoppers finish on a payment provider’s page and never come back. Shopify says in its help center that analytics tools and admin numbers will not match. On general-audience sites, ad blockers hide roughly a tenth of visitors (Quantable’s measurements, as viewed in November 2026).
What matters is that this usual gap stays at roughly the same level every week.
| What happened | GA4 purchases | Orders | Ratio |
|---|---|---|---|
| Fewer customers | Down | Down by the same amount | Stays at its usual level |
| Tracking broke | Down | Unchanged | Steps down from one week on |
The week where the ratio steps down is the week something broke. If it slides gradually instead, that looks less like a breakage and more like a slowly growing share of visits going uncounted.
One caveat: leave the most recent days out. Google notes in its official explainer that GA4 can take up to 72 hours to process events (as viewed in November 2026). A ratio taken partway through the current week can look low even when nothing is wrong. Drop the last three days and compare only completed weeks.
Check: lay out eight to ten weeks of the ratio and get a feel for its usual level first. Then look only at the weeks that fall outside it.
3. What changed in the week the ratio dropped?
Once you know the broken week, think about what changed on your store that week. A purchase is only counted when the tracking tag fires on the order confirmation page. If anything along the way changes, orders can keep arriving while purchases stop being counted.
Typical candidates include:
- A theme update or a new theme
- Changes to checkout or payment methods
- A replaced order confirmation page
- Swapped tracking tags, or a new way of managing tags
- A new consent banner, or changes to its settings
The cause differs from store to store. These are examples, not a diagnosis. Match the week the ratio dropped against your change log and your platform’s update history, and check the closest candidates one at a time.
Check: does the week the ratio dropped line up with the date of a change? If not, consider a change from the previous week that only took effect after the weekend.
4. GA4 cannot give you this ratio on its own
There is one thing missing from all of this. The denominator, your order count, is not in GA4. GA4 only knows the purchases its tag counted. When the tag breaks, GA4 has no record that it broke.
That means as long as you only look at GA4, you cannot decide on the spot whether you stopped selling or stopped counting. And without that decision, you end up reworking ads and product pages when the problem is tracking, or doubting your tracking while sales are genuinely falling.
To have this ratio, you need either to copy your weekly order count into a sheet, or a setup that puts order data and GA4 purchases side by side. What is missing is not effort. It is the information you need to decide.
So you do not have to investigate from scratch every time
The order to follow when purchases suddenly drop:
- Check admin orders for the same dates and time zone.
- Lay out GA4 purchases ÷ orders by week and look for a step. Leave out the most recent days.
- Match the step week against what you changed.
- After fixing it, confirm that the next week’s ratio is back at its usual level. If not, something is still being missed.
Once the ratio is back to normal and you trust your tracking again, return to the numbers you watch every week. We cover which ones in The only 5 numbers an ecommerce store needs to check in GA4 each week.
Related articles
- November 20, 2026The traffic source saved on every WooCommerce order — how to read it and where it stops
WooCommerce records where each customer came from on the order itself. Line those order sources up by channel and the order records tell you which channels are producing sales. But only the last channel is kept, and visits that did not end in an order leave nothing behind. Based on the official documentation, this post covers how to read it and what to decide with it when it disagrees with GA4.
- November 18, 2026When does the second order come — use your own orders to pick the day you follow up
If you send your follow-up email or coupon a set number of days after the first purchase because another store does it that way, you are probably guessing. When the second order arrives depends on what you sell. Count the days from first to second order in your own data, send one message just before the peak, and judge it by comparing groups of customers who first bought in different months.