How duplicate events happen
A duplicate is one action reaching the platform as several. The usual source is two things sending the same event without knowing about each other: a tag placed directly in the theme alongside the same tag in a container, an ecommerce plugin with built-in tracking plus a manual setup, or two containers left behind after a redesign.
Triggers cause the rest. A conversion tied to a thank-you page view fires again every time the visitor refreshes, returns with the back button or bookmarks the page. A form event set to fire on both submission and the following page view counts twice by design. Single-page sites can fire on every internal route change. And when a browser pixel and a server-side sending method both report the same purchase, they are counted separately unless each carries a shared identifier the platform can match them on.
Why duplicate events matter
The reporting damage is obvious: conversions look higher than they are, cost per lead looks lower, and revenue in analytics disagrees with the accounts. Decisions made on inflated numbers tend to be expansive ones, so budget grows on the basis of results that never existed.
The costlier damage is invisible. Automated bidding learns from conversion signals, so if certain paths duplicate and others do not — a checkout that refreshes, a landing page with a legacy tag — the platform concludes those journeys are more valuable and buys more of that traffic. The account then drifts towards whatever is easiest to double-count, which is not the same as whatever is most profitable.
Common mistakes with duplicate events
The first is trusting deduplication to happen on its own. Matching only works when both sends carry the same identifier for the same action, in a field the platform is actually reading; without it, two reports of one purchase are two purchases. The second is leaving the counting setting wrong: an enquiry form should generally count once per click, since the same person filling in two forms is still one lead, while sales usually count every occurrence.
The third is checking totals rather than journeys. Duplication is easy to see when you complete one conversion and watch the event count, and almost impossible to see in a monthly total that merely looks encouraging. The fourth is removing the wrong copy — deleting the container tag while a hardcoded snippet survives leaves you with a tag nobody can maintain.
How to act on it
Test the way a real visitor behaves. Complete a conversion, refresh the confirmation page, press back, and submit a second form; then check what was recorded. Anything more than the number of genuine actions is a duplicate to chase.
Fix the cause rather than the symptom: remove the redundant tag, trigger on the submission event rather than a page that can be reloaded, and make sure every purchase carries an order identifier so repeated sends collapse into one. Where the same event is deliberately sent from both the browser and a server, follow the platform’s matching requirements exactly, as described under event deduplication. Then confirm the repair with the same walk-through you would use for any tag validation, on a phone as well as a desktop.
