How to investigate duplicate form conversion counts

01

Identify what each system counts

Write down the form action, its conversion event and the sales record it creates. In Google Ads, the count setting applies to each conversion action: 'every' counts every conversion after an interaction, while 'one' counts one per interaction for that action. Neither setting is a universal rule for unique people. Multiple conversion actions can also appear in a report.

Choose a closed period and preserve the original event timestamps and source, along with available event identifiers. Compare them with form and sales records using identifiers your privacy rules permit. Note which system uses event date and which attributes activity to an earlier interaction. A platform may include modeled conversions that cannot be matched to individual records.

Inspect test submissions separately from customer inquiries. Restrict access to contact details, and check platform requirements and applicable privacy rules before sharing identifiers or uploading sales data.

02

Trace one submission

Hypothetically, one visitor submits a form, sees the confirmation page, then refreshes it. The website log shows one submission but two conversion events. The sales system has one inquiry. That pattern suggests the confirmation page may fire the event again. Inspect the event implementation and browser behavior before treating the second event as a duplicate.

Now consider a different case: the same person sends a second form for a separate service. Two submissions may be legitimate even if a single contact record holds both. Decide whether your reporting unit is a submission, opportunity or person, and preserve both requests for the team to handle.

Compare counts at each step: submissions, events, distinct inquiry records and unmatched items. List possible causes without forcing a match. Missing identifiers, consent choices, attribution rules and reporting delays can leave a remainder that cannot be explained from the available evidence.

03

Fix the confirmed cause

If a refresh fires an unintended event, fix the trigger at the point it occurs and document the change. If the issue is the intended counting rule, review whether 'one' or 'every' fits the business event. Google states that changes apply to future reporting. Do not assume a settings change repairs historical counts.

After a change, compare a small set of permitted test submissions with event and sales records. Remove or label test entries according to your reporting procedure. Keep a reconciliation log with the date, affected event and remaining unmatched counts. A lower conversion count is not evidence of better lead quality by itself.