How to troubleshoot inaccurate conversion tracking for social media advertising?

Conversions appear in the ad dashboard, but the corresponding inquiries cannot be found on the website; or orders are clearly increasing, while the ad platform reports almost no conversions. When this happens, do not rush to adjust the budget or reinstall the tracking code. The discrepancy may stem from differences in measurement methods, missing or duplicate event reporting, or changes to attribution settings. Each cause requires a different response, and making changes hastily can make subsequent data harder to compare.

When troubleshooting, start by selecting one specific conversion action, such as a successful form submission or a completed order payment. Then check each step in sequence: “user completes the action, website records it, event is sent, platform receives it, and ad attribution is applied.” Identify where the discrepancy occurs before deciding whether the marketing team, website developer, or ad account administrator should handle it.

First, confirm that you are comparing the same thing

Figures in ad platforms, website analytics tools, and CRM systems may not align directly. An ad platform may attribute conversions based on when an ad interaction occurred, while the website or CRM records them based on when a form was submitted. A platform may include conversions attributed to both ad clicks and views, whereas a CRM records leads that actually entered the system. Even when tracking is working properly, the figures may differ.

Before checking, standardize the date range, time zone, and conversion definition. B2B websites should distinguish between “clicking the submit button,” “successful form submission,” and “qualified inquiry.” The first may include attempts that failed validation, while the last requires further screening based on manual review or business rules. Cross-border e-commerce stores should also distinguish between orders created, payments completed, and orders subsequently refunded to avoid comparing data from different stages.

Also check the attribution window used by the ad platform and whether the report includes view-through conversions. Recent data may change due to processing delays or subsequent attribution adjustments, so use data that has had sufficient time to process wherever possible. If you cannot confirm that the measurement methods match, first record each system’s settings, then determine whether the discrepancy is abnormal.

Follow a real user journey to locate where the data flow breaks

Choose a controlled test environment or test order, start from the ad landing page, and complete the target action along the user journey. Before testing, record the time, page URL, device, and browser. When submitting a form, use identifiable test information and notify the sales or order-processing team in advance to prevent test data from entering the normal follow-up process.

You do not need to click a live ad for every test. To check whether a website event fires, you can first visit the page directly and use the platform’s testing tools. Verifying whether ad attribution works is a separate step and should be assessed using the platform’s supported testing methods and the actual ad interaction conditions. A single test that is not attributed to an ad does not, by itself, prove that the tracking code is not working.

• No record on the website: Check form validation, API requests, payment result callbacks, and redirect pages to confirm whether the user’s action actually succeeded.
• Website has a record, but the platform’s testing tool shows no event: Focus on code loading, event trigger conditions, data transmission, and user consent status.
• The platform received the event, but the ad report shows no conversion: Check the conversion event selected for the campaign, attribution conditions, test traffic source, and data processing time.
• The platform reports significantly more conversions than the website: Check for duplicate triggers, overly broad event definitions, and differences in measurement methods such as view-through attribution.

An event appearing in a testing tool only confirms that it reached the relevant stage during that particular test. You should also confirm that it came from the correct data source, that the event name matches the ad optimization objective, and that it works as expected under real-user consent conditions.

Correctly installed code does not mean events are configured correctly

After a website redesign, theme change, form plugin adjustment, or domain migration, tracking code may be omitted or both the old and new versions may remain active. First, use the platform’s event diagnostics, browser developer tools, or tag management preview mode to check whether the expected data source loads on the target page and which events are sent when the action is completed.

A common mistake with form inquiries is treating a button click as a successful submission. Users who have not completed required fields, API errors, and even repeated button clicks can all inflate the figures. A more reliable approach is to trigger the event only after the website confirms a successful submission. If you use a thank-you page, also account for duplicate records caused by users refreshing or directly visiting that page.

For e-commerce orders, check the trigger timing and order status. Treating a visit to the checkout page as a completed purchase will overstate sales; relying only on the page users return to after payment may miss orders when they do not successfully return to the website. If payment results are confirmed by the website backend, consider sending the corresponding event server-side and validating it against the order status.

When using both browser-side and server-side tracking, the two channels may report the same conversion. Check the platform’s deduplication requirements, such as whether the event name and event identifier match. Field requirements vary by platform, so do not assume that the platform will deduplicate events simply because they occurred close together. For multilingual websites, check every language version and subdomain to ensure tracking is fully installed beyond the main site.

Even after an event is received, check the attribution path

Conversions from social platform ads typically require the platform to associate an event with a previous ad interaction. Cross-device visits, different browsers, third-party payment pages, and browser privacy restrictions can all make this association more difficult. A platform’s failure to attribute an order does not mean that the order was unaffected by advertising; conversely, conversions reported by a platform are not the same as incremental sales generated exclusively by ads.

When checking landing page redirects, look for ad parameters that may be lost during redirects, cross-domain navigation, or payment round trips. If tracking is configured through a tag management tool, also check that the trigger rules cover the domains and page paths actually visited. For server-side events, the matching information passed must comply with platform requirements and applicable privacy rules. If users have not given consent or have declined tracking, follow the corresponding settings; do not bypass consent mechanisms to make the figures match.

Finally, return to the account settings and confirm which event the campaign is optimizing for, whether the event comes from the correct website data source, and whether anyone has recently changed the attribution settings. After fixing the issue, record the change time and test results, then monitor subsequent data using the same measurement methods. For discrepancies that persist, separately record actual website sales, events received by the platform, and conversions attributed by the platform. This is more useful for evaluating the budget than trying to force all three figures to match.