Duplicate conversions in Google Ads: how to find, fix and date the cause

What are duplicate conversions in Google Ads?

Duplicate conversions are extra conversions recorded for an action that happened only once. One purchase becomes two purchases, one lead becomes three leads, and every metric built on conversions moves with them: conversion rate goes up, CPA goes down and ROAS looks better than it is.

The cost is bigger than a wrong report. Primary conversion actions are the ones reported in the "Conversions" column and used for bidding, according to Google's help page on primary and secondary conversion actions. If those numbers are inflated, Smart Bidding learns from conversions that never happened and can push budget toward the wrong campaigns, keywords or audiences.

The typical symptom is a sudden rise in conversions that your source of truth does not confirm. Example: Google Ads reports 84 purchases for last week while your store backend shows 46 orders, and the gap appeared on one specific day rather than growing slowly. Some gap is normal (attribution, conversion date versus order date, consent). A gap that jumps overnight usually means something in the tracking changed.

What causes Google Ads conversions to be counted twice?

Almost every case comes down to the same event reaching Google Ads through two paths, or one path firing twice. The table lists the causes you will see most often and how each one shows up.

Search Engine Land's guide to duplicate conversions adds a common e-commerce case: a store's built-in Google channel or a checkout app adds its own tracking on top of a custom GTM setup, so both fire on the same order.

Two of these causes are settings, not code. Google describes "Every" as counting every conversion after an ad interaction and "One" as counting only one conversion per ad click, and recommends "One" for leads, where one unique lead per click is what adds value (Google Ads Help).

How do I check whether my Google Ads conversions are duplicated?

Compare Google Ads with a source that counts each real action once, then trace the gap to one conversion action and one tag. Work in this order:

1. Pick a source of truth. Orders in your store or payment system for purchases, contacts or deals in your CRM for leads. Export daily counts for the last 30 to 60 days.
2. Compare by day. Put Google Ads conversions next to the backend count. Look for the day the ratio changed, not for a perfect match.
3. Split by conversion action. Segment the Google Ads report by conversion action. If one action carries the whole jump, you know where to look.
4. List your actions. Open Goals > Conversions > Summary and check which actions are Primary. Two primary actions that describe the same event, for example "Purchase" from the Google Ads tag and "purchase" imported from GA4, both feed the "Conversions" column.
5. Check the count setting. Open each lead action and confirm whether "Count" is set to "One" or "Every".
6. Watch the tag fire. Place a test order or submit a test form with Google Tag Assistant open and count the Google Ads conversion hits. In Chrome DevTools, filter the Network tab for googleadservices.com/pagead/conversion; Google notes that the transaction ID appears in the payload as oid.
7. Reload the confirmation page. If a refresh sends a second hit, reloads are inflating your purchases.

Steps 1 to 3 tell you whether you have a problem and when it started. Steps 4 to 7 tell you which mechanism causes it. Keep the test order numbers so you can match them to hits later.

How do I fix duplicate conversions in Google Ads?

Remove the second path first, then add deduplication so a single path cannot count the same action twice. Match the fix to the cause you found.

Add a transaction ID to purchase conversions

When two conversions for the same conversion action carry the same transaction ID, Google Ads treats the second one as a duplicate and does not count it. Google's rules for the ID (source):

• Unique per transaction and generated dynamically by your backend or e-commerce platform, never hard-coded.
• Up to 64 characters, with no personally identifiable information, URLs, email addresses, prices or phone numbers.
• Deduplication works within one conversion action, so it will not stop two different actions from counting the same order.
• It does not remove duplicate view-through conversions, and transaction IDs are not shown in Google Ads reports.

In Google Tag Manager the usual setup is a data layer push with the order ID, a Data Layer Variable named transaction_id, and that variable mapped in the Google Ads conversion tag.

Keep one primary action per real event

If the Google Ads tag and a GA4 import track the same purchase, keep one as Primary and set the other to "Secondary (observe only)". Secondary actions stay visible in the "All conversions" column but are not used for bidding, unless you add them to a custom goal.

Use "One" for leads

Switch lead actions to "One" in Goals > Conversions > Summary > the action name > Edit settings > Count. Google states the change only applies to future conversion reporting; past data keeps the old count.

Clean up the tags

Remove hard-coded tags that duplicate GTM, turn off the extra tracking in platform plugins, and tighten triggers so they match one event. In GTM's Advanced Settings, "Once per event" is the default tag firing option. "Once per page" stops a tag from firing more than once per page load, but a reload creates a new page load, so it does not replace a transaction ID.

Correct past data if it matters

Conversion adjustments let you retract or restate conversions that were tracked with an order ID, or with the GCLID plus conversion time. Google says adjustments can be made within 54 days and recommends uploading them at least 24 hours after the conversion (source).

How do you find the change that started the duplicate conversions?

Duplicates almost always start on a specific date, and that date matches a change someone made. Find the first day the conversion-to-order ratio moved, then list every change from the days before it.

1. Put everything on one timeline. GTM container versions published, website or theme releases, new plugins or checkout apps, conversion actions created or edited, GA4 imports, goal changes and changes made by platform automations.
2. Rule out tracking first. If conversions rose while clicks, sessions and backend orders stayed flat, treat it as a tracking change before you credit a bid strategy or a new ad.
3. Test one hypothesis at a time. Example: the jump started on 14 September, and GTM version 37, published on 13 September, added a second trigger to the purchase tag. Pause that trigger in a test, place an order and count the hits. If the duplicate disappears, you have the cause.
4. Check what else changed that day. A sale, a new campaign or a budget increase can explain part of the rise, so separate the real lift from the tracking error.

Google Ads change history shows edits inside the account, such as new conversion actions or goal changes, but not GTM publishes, site releases or plugin installs. Those live in other tools or in someone's memory, which is why a shared marketing change log saves hours here. The opposite problem, a tracking change that hides real conversions, is covered in our post on the consent mode conversion drop.

What should you log after fixing duplicate conversions?

Log the fix as a tracking change with an exact date, because the fix itself will move your metrics. Once duplicates stop, reported conversions fall, CPA rises and ROAS drops, even though the business did not change.

• When the duplication started and when it was fixed (date and time).
• Which conversion actions were affected and roughly by how much (example: "Purchase" counted about 1.8 conversions per order).
• What changed: transaction ID added, action set to Secondary, count switched to "One", trigger removed.
• Who made the change and where (GTM version, Google Ads, website release).
• Whether past data was retracted with conversion adjustments or left as it was.

Add the same dates as annotations in GA4 and in your client report, so nobody reads the drop after the fix as a performance problem. If the client asks why results fell, our guide on how to explain a conversion drop to a client gives you a structure. If you later see a real fall in conversions, start from the method in why did my conversions drop.

Changeline keeps these entries on one to-scale timeline per client, next to Google Ads, Meta Ads, website and CRM changes, so the next person who sees the jump or the drop finds the reason quickly. You can start a change log for free while Changeline is in early access.