The Most Common Google Ads Double-Install on Shopify

If your Google Ads conversions look a little too good, there is a boring reason it happens more than anything else on Shopify: the same purchase gets counted twice. Once by your theme, once by an app. Nobody set out to break anything. Two things just ended up doing the same job, and Google has no idea they are the same sale.

This is the double-install. It is quiet, it is common, and it makes your ads look better than they are.

How the double-install happens

Here is the usual story. Someone sets up Google Ads conversion tracking early on. Maybe they paste the tag into the theme code, or drop it in the Shopify checkout settings, or add it through Google Tag Manager. That works. Sales start showing up in Google Ads.

Months later, a different person, or the same person on a different day, installs a tracking app to "get better data." Plenty of these apps exist. They promise cleaner conversion tracking, server-side firing, better match rates. So they install it, connect it to Google Ads, and walk away.

Now two things fire on every order:

• The original tag baked into the theme or checkout
• The new app, doing exactly what it was installed to do

Both report the purchase to the same Google Ads conversion action. Google receives two pings for one order and counts two conversions. Your numbers just doubled, and not because you sold more.

Why Shopify makes this easy to miss

On most platforms you would trip over the duplicate code. On Shopify, the two sources usually live in completely different places, so you never see them side by side.

The theme tag hides in your Liquid files or in the "Additional scripts" box on the checkout page. The app tag lives inside the app, which you almost never open once it is running. One is code you can read. The other is a black box with a settings screen.

So when you look at your theme, you see one tag and think you are fine. When you look at the app, you see it is "connected" and think you are fine. Neither view shows you the other. The overlap only exists in Google Ads, where both are quietly adding up.

There is also the checkout change working against you. Shopify moved a lot of stores toward its newer checkout and "Customer events" (also called Web Pixels), where tracking is supposed to run through a managed pixel instead of raw script in the checkout. During that transition, plenty of stores kept the old theme script AND added the new pixel-based method AND kept an app. That is three chances to count the same sale.

The numbers stop making sense

You do not need to be an analyst to catch this. You need a calculator and thirty seconds.

Pull up your Google Ads conversions for the last 30 days. Then pull up your actual paid orders for the same window from Shopify. Not total orders. Orders you can reasonably tie to Google Ads. Now compare.

If Google Ads is reporting roughly twice the orders you actually got, you have a double-install until proven otherwise. Some tells:

• Conversions that sit at a suspiciously clean 2x your real order count
• A conversion rate that feels too high for your traffic
• Revenue in Google Ads that is close to double what Shopify says you made
• Two separate "purchase" or "conversion" entries in your Google Ads conversion actions list, sometimes with slightly different names

That last one is the fastest check. Go to Google Ads, open Goals, then Conversions. Look at the list of conversion actions. If you see two things that both mean "someone bought," and both are set to count toward your bidding, that is your double-install staring back at you.

Why this is worse than a reporting typo

A double-count is not just an ego problem in a dashboard. Google Ads uses those conversions to decide where to spend your money. Smart Bidding, Target ROAS, Maximize Conversions: all of them optimize toward the conversions you feed them.

If you feed Google two sales for every one real sale, it believes some clicks and keywords perform twice as well as they do. It leans into them. It pays more for them. You are now bidding on a fantasy version of your own performance, and the real winners get less budget than they should.

You also cannot trust a single decision made from that data. Every "this campaign is crushing it" call, every "let's scale this," every pause of a "loser" is built on numbers that are off by a factor you cannot see.

We learned this the hard way with our own store, Outer Envy. The tracking there once reported 67,184 conversions against about 900 real orders. It showed a 145% conversion rate, which is not a thing that can physically happen, and a cost per purchase of $0.06. We optimized toward those phantom events for 16 months and burned around $3,200 a year in paid search chasing them. After the fix, reporting matched reality to the cent: $269.90 reported equaled $269.90 actual. A double-install is a smaller version of that same disease.

How to fix it without breaking everything

Do not rip out both. That is how you end up with zero tracking. Pick one source of truth and turn off the other.

1. List every source. Check the theme code, the checkout "Additional scripts," Shopify Customer events, Google Tag Manager, and every installed tracking app. Write down anything that fires a Google Ads purchase.
2. Decide who wins. Usually you keep the cleanest, most modern method and retire the rest. If an app is doing server-side tracking well, keep it and remove the theme script, or the reverse. One purchase conversion, one source.
3. Turn the other off, do not just ignore it. A tag left in the theme keeps firing even after you stop looking at it.
4. Test one real order. Buy something, then watch Google Ads. One order should produce one conversion. If you still see two, you missed a source.

The goal is simple. One sale, one count, everywhere.

Not sure which of your tags is the twin?

If your Google Ads numbers feel too high and you cannot tell whether your theme, your checkout, or an app is the one double-counting, that is exactly what the AdProof Audit is for. We look at what your tracking is actually counting, touch nothing, and hand you a plain-English report plus a short recorded walkthrough. No pressure. Just a clear picture of whether your sales are being counted once or twice.