How to Use Google Ads Data Exclusions After a Clinic Conversion-Tracking Outage

A Google Ads data exclusion can protect Smart Bidding from a verified conversion-tracking outage. It does not repair the tag, restore missing clinic records, or remove data from reports. The exclusion tells eligible bidding strategies to disregard affected click data when learning and setting bids.

The date range must follow clicks, not only the hours when conversions stopped arriving. Include earlier clicks whose conversions would normally have appeared during the outage. Apply the narrowest accurate campaign and device scope, preserve the incident evidence, and treat reporting backfill as a separate decision.

When does a clinic conversion outage qualify?

Google recommends data exclusions for conversion-data outages, including broken tracking, an incorrectly placed tag, a website outage, or a temporary failure to upload offline conversions. In each case, inaccurate conversion or conversion-value data could distort conversion-based Smart Bidding.

A sudden decline in appointments does not qualify by itself. Neither does a holiday, a normal change in demand, a small sample, or a disputed lead-quality definition. First prove that the measurement system reported incorrect data for a bounded period.

Keep the incident proof outside Google Ads: alert time, affected conversion actions, first bad observation, last known good event, repair time, missing or inflated records, affected devices, and affected campaigns. The conversion audit proof guide helps separate configuration, processing, reporting, and attribution evidence.

What does a data exclusion change?

A data exclusion changes the conversion and conversion-value data that Smart Bidding uses. Google says the tool is available only for conversion-based and value-based Smart Bidding strategies. It supports Search, Display, Shopping, and Performance Max campaigns, while Hotel and Travel campaigns are not supported.

The exclusion does not alter the Conversions columns or erase affected events from reports. Google states that excluded conversions remain visible in reporting. A report can therefore show conversions from the period even though Smart Bidding no longer uses the associated click data.

Describe the action accurately in an incident log: “Excluded affected clicks from Smart Bidding inputs.” Do not write “deleted bad conversions,” “fixed reporting,” or “restored attribution.” Those outcomes are outside what the feature does.

Why must the date range follow clicks and conversion delay?

Google applies data exclusions to clicks. When a click is excluded, conversions associated with it are also excluded from Smart Bidding. A conversion recorded during an outage may belong to a click from an earlier day, so selecting only the visible outage dates can miss affected learning data.

Measure the normal delay between click and reported conversion for the affected action. Then move the exclusion start back far enough to cover clicks whose conversions would have arrived during the broken period. Google recommends covering at least 90% of clicks associated with the affected conversion data.

Google gives a concrete example: if uploads failed from October 15 through October 18 and the conversion delay is five days, the exclusion should cover clicks from October 10 through October 18. Use the clinic account’s observed delay and incident window instead of copying the example dates.

The clinic funnel report by original lead cohort provides a related reporting principle. Later outcomes belong to the acquisition cohort that produced them. The exclusion uses the same temporal discipline because bidding connects reported conversions back to clicks.

How narrow should the exclusion scope be?

Scope should match the proven fault. Google’s setup allows selection of campaigns and devices, or an account-wide exclusion when the evidence supports it. Exclusions can be applied at manager-account or sub-account level. Start and end times use the account’s time zone.

Incident evidence	Candidate scope	Do not assume
One mobile tag failed	Affected campaigns and mobile devices	Desktop data was also wrong
One campaign used a broken page	That campaign and proven dates	Every account campaign was affected
Account-wide import stopped	Campaigns using the affected imported action	Unrelated conversion actions failed
Scope remains unknown	Continue investigation	A broad exclusion is safer

A broad exclusion discards more learning data. A narrow exclusion that misses affected clicks leaves bad data in the bidder. Choose the smallest scope supported by incident evidence, which may not be the smallest selectable option.

How should the team apply the exclusion?

Google’s current workflow places data exclusions under Adjustments in Google Ads. Before creating one, record the exact dates, times, time zone, campaigns, devices, conversion actions, and delay assumption. Google warns that exclusions can change bid-strategy performance, so the final review should happen before creation.

1. Confirm the conversion data is wrong and identify the root measurement fault.
2. Repair or contain the fault so the outage has a defensible end.
3. Measure the normal click-to-conversion delay for the affected action.
4. Translate the outage into an affected click-date range.
5. Select only the proven campaigns and devices.
6. Verify the account time zone and manager or sub-account scope.
7. Save the evidence and approval before creating the exclusion.
8. Monitor bidding and spend while the strategy stabilizes.

Apply the exclusion quickly once the scope is known. Google says planned outages can be configured in advance and typically take effect within a day. Do not use that guidance to guess at an uncertain incident. A wrong future window still sends the bidder inaccurate instructions.

What happens after the exclusion is applied?

Google advises against removing a data exclusion after application. Past-date exclusions should begin to stabilize after a few days, but an outage affecting a week or more of clicks can produce fluctuations for one or two conversion cycles.

Watch spend, bids, conversion volume, and conversion value without declaring immediate recovery. Google notes that bids and spending can fall temporarily. It also suggests keeping budgets near historical spend for a few cycles and adjusting CPA or ROAS targets when necessary. Any such change needs its own approval and rollback rule.

The value-based bidding readiness checklist explains why delay, mature observation windows, goals, and data quality still matter after the platform accepts a setting. An exclusion reduces one source of bad learning data. It does not make the remaining conversion feed accurate.

Should missing conversions be backfilled?

Google generally advises against backfilling data after an outage because late conversions can affect Search and Shopping bidding performance. If a data exclusion was applied, Google allows a reporting-only backfill after waiting at least five days. It says not to remove the exclusion.

Keep the two purposes separate. The exclusion protects bidding from affected clicks. A backfill may restore reporting completeness. It does not reverse the exclusion or recreate the learning sequence that would have existed if tracking had worked on time.

Before any backfill, reconcile source records, remove duplicates, preserve original conversion times, and document why reporting needs the late events. The offline conversion outbox report shows how to keep local event state, Google request state, Ads reporting, and attribution in separate evidence layers.

What should the incident record contain?

A useful record lets another operator reconstruct the decision without reading patient or appointment data. Include the incident ID, detection time, root cause, repair time, affected conversion actions, account time zone, measured delay, click-date range, campaign and device scope, approval, exclusion identifier, and monitoring window.

Record explicit non-claims as well:

• the exclusion did not change Ads reporting;
• the exclusion did not repair the source tracking system;
• the exclusion did not prove attribution or business recovery;
• any backfill was reporting-only and happened after the documented wait;
• patient details and clinical data were not included in the incident artifact.

Review the incident after one or two observed conversion cycles when the affected period was long. Retire temporary target or budget adjustments through a separate approved decision. Do not remove the data exclusion to make a dashboard look consistent.

Frequently asked questions

Do data exclusions remove conversions from reports?

No. Google says data exclusions affect the data used by Smart Bidding, not conversion reporting. The excluded period can remain visible in the account’s conversion columns.

Should the exclusion cover conversion dates or click dates?

Use affected click dates. Extend the start date to account for normal conversion delay, because conversions reported during the outage may be attributed to earlier clicks.

Can the team remove an exclusion after fixing the data?

Google advises against removing an applied exclusion. Preserve it, document the repair separately, and let the strategy stabilize on accurate post-incident data.

Can missing conversions be uploaded later?

Google generally discourages backfills. When an exclusion was applied, a reporting-only backfill can occur after at least five days, without removing the exclusion. Review duplicates and original times first.

References

• Google Ads Help, “About data exclusions,” retrieved 2026-08-16, https://support.google.com/google-ads/answer/10370710?hl=en
• Google Ads Help, “Use data exclusions for conversion data outages,” retrieved 2026-08-16, https://support.google.com/google-ads/answer/10276486?hl=en