Something changed last week. Traffic looks normal, orders do not. Everyone is now asking what you did to the site, and the obvious instinct is to start undoing things.
Do not. A sudden drop has a small number of likely causes, and they are checkable in a specific order. Most of them are not the store. Two of them are not even real.
What follows is the triage we run, in order. Each step takes minutes, and you stop at the first one that explains the drop.
Step 1: check that the drop is real
Two stores in three that come to us with a sudden drop have one of these, and no drop at all.
The comparison window is doing the work. Last week had a promo, a newsletter, a payday, a holiday. This week does not. The rate fell because the previous period was unusually good, not because this one is bad. Compare against several previous periods, not just the one immediately before, and against the same period last year if your business is seasonal.
The numbers are too small to mean anything. On a store with a few hundred orders a month, the conversion rate swings by a meaningful-looking amount week to week for no reason at all. A day with two orders instead of five is not a trend. Widen the window until the line stops jumping, and look at that.
If the drop survives both checks, it is real. Keep going.
Step 2: check whether measurement broke
This is the most common real cause of a sudden drop, and almost every article on the subject puts it last or leaves it out. A tracking change looks exactly like a conversion collapse: traffic unchanged, conversions gone.
What to check, in this order:
Did anything deploy? Compare the date of the drop against your theme deploy history. A theme push that drops a tracking snippet, or reorders scripts so one no longer fires, will do this instantly. Shopify keeps the history, so this is a two-minute check and it is the single highest-yield one.
Did an app update or get installed? Apps inject script. An update that changes how or when it loads can break another app's tracking, and a consent app that started blocking analytics by default will zero out your reporting without touching a single order.
Did your consent banner change? A switch from implied to explicit consent, or a new region added, will cut measured conversions sharply while actual orders stay flat. This one catches a lot of stores and it is not a bug, it is the banner doing its job.
Are you comparing the right two numbers? Shopify's own conversion rate and a GA4 conversion rate are built differently and will never match. If your drop appears in one and not the other, the drop is in the measurement, not the store.
Did something start double-counting, then stop? A duplicate tag inflating conversions for weeks, then removed, produces a dramatic drop from a number that was never true.
The test that settles it: look at actual orders in Shopify admin for the period. If order count is flat and the rate fell, you have a measurement problem. If orders genuinely fell, keep going.
Step 3: check whether the traffic changed
The rate is a fraction. A drop can come entirely from the denominator growing with people who were never going to buy.
A new campaign or channel started. Cold paid social, a new influencer, a broad-match campaign. Volume up, intent down, rate down. Nothing is broken. Split conversion rate by channel and the culprit shows up immediately.
A ranking changed. Picking up a high-volume informational query sends a lot of readers who are not shoppers. Check Search Console for new queries and landing pages in the period.
Bot or referral traffic spiked. A scraper or a spam referral inflates sessions and nothing else. Look for a sudden lift in sessions with near-zero engagement, from one source or one country.
A device or geography shifted. If mobile's share of traffic jumped and mobile converts worse on your store, the blended rate falls while neither segment changed.
In all four cases the fix is not on the product page. If the rate is stable inside each channel and only the blend moved, you are done investigating.
Step 4: check whether the store changed
Now we are looking at the store itself, and the method is the same as step 2: line up dates.
A theme deploy. Any change to the cart, product template or checkout-adjacent code. Look at what shipped, not at whether it was supposed to be harmless.
An app installed, updated or removed. Especially anything touching the cart, checkout, shipping rates, reviews or popups. An app removed can be as damaging as one added, because uninstalling does not always clean up the code it left behind.
A payment method disappeared. Expired credentials, a provider dropping support, a gateway disabled during a billing problem. If one of your main payment options stopped appearing, a chunk of buyers simply cannot pay. This shows up as checkout started but not completed.
Shipping rates or zones changed. A rate table edit that accidentally removes a zone, or raises a cost sharply, hits the cart and checkout steps hard and leaves everything above them untouched.
Something went out of stock. If your top seller is unavailable, the rate falls and nothing is broken. Check whether traffic is still arriving on pages for products you cannot sell.
A price change. Yours, or a competitor's promo that changed what your price looks like next to theirs.
A discount code expired. If a code was circulating widely, its expiry removes a conversion driver overnight.
Step 5: check whether something outside changed
Last, because it is least actionable and least likely, but it does happen.
A payment provider outage. A Shopify platform incident. A browser or OS update changing how something renders or how cookies behave. A marketplace or competitor running an aggressive promotion in your category. Seasonality that you have not been in business long enough to recognise yet.
Shopify publishes a status page, and your payment provider does too. Both are worth a look before you rebuild anything.
The one-hour triage, in order
If you do nothing else from this article, do this, in this sequence:
- Pull order count, not conversion rate, for the period and the three before it. Did orders actually fall?
- Check your theme deploy history and app activity against the date of the drop
- Split conversion rate by channel. Is it falling everywhere, or in one place?
- Split by device. Is it mobile only?
- Walk the funnel: sessions, add to cart, checkout started, orders. Which step lost the most?
- Place a test order yourself, on a phone, with a real payment method
- Check Shopify status and your payment provider's status
Steps 1 and 2 resolve most cases. Step 6 catches the ones that nothing in analytics will show you.
If the drop is real and it is the store
Then you are no longer dealing with a sudden drop, you are dealing with a funnel problem, and the step that lost the most in triage point 5 tells you where. Each gap has a different set of fixes, which is a longer subject: we wrote about finding and fixing the step that loses buyers separately.
The short version of where to look:
People adding to cart and leaving points at the cart. People reaching payment and stopping points at checkout. People leaving before they ever see a product often points at load time.
How to not be here again
A sudden drop is cheap to diagnose and expensive to notice late. Three things make the difference:
Watch orders, not just the rate. A rate can fall for reasons that do not cost you money. Order count cannot.
Alert on the funnel, not the total. Set something that tells you when checkouts started falls sharply, or when add-to-carts fall, rather than waiting for the monthly revenue number to be wrong.
Keep a change log. Date, what changed, who changed it: deploys, app installs, rate edits, price changes, campaign launches. Thirty seconds per entry, and it turns this entire article into a two-minute lookup. Most of the time we spend diagnosing a sudden drop is spent reconstructing what changed and when, from memory, after the fact.
Common questions
In order of likelihood: the comparison period was unusually good, tracking broke, the traffic mix changed, or something changed in the store. Check order count first - if orders are flat and only the rate fell, you have a measurement problem rather than a conversion problem.
Compare order count in Shopify admin against the same period before. Shopify's order data does not depend on your analytics setup. If orders held steady, the drop is in the reporting.
Not necessarily, and often not at all. Conversion rate is a fraction, so adding lower-intent traffic lowers it even when revenue rises. Look at orders and revenue alongside the rate, and split the rate by channel.
It depends on your volume. A store with a few hundred orders a month will swing several tenths of a percentage point week to week with nothing wrong. Widen the window until the line steadies, then judge.
Yes, and it is one of the most common real causes. A deploy that drops a tracking script, breaks a cart interaction on one device, or changes how shipping rates display will do it immediately. Comparing the drop date against deploy history is the highest-yield check available to you.
Only if you know what to roll back to and why. Blind reverting destroys the evidence you need and often reintroduces something else. Diagnose first, which in most cases takes under an hour.
This is the triage we run before quoting any conversion work, because about half the time it ends with telling the client nothing is broken.
If you want it run properly on your store, our paid conversion audit is $900, takes five working days, and the findings are yours whether you build with us or not. Our Shopify CRO services page covers what happens after.
