Google Ads updates Conversion Lift setup guide with stricter tracking requirements

The Conversion Lift documentation, updated on November 13, 2026, did not change the incrementality test's methodology, but it tightened the technical tracking requirements that feed into the result. The update itself now explicitly lists unsupported configurations, and Google's own recommendation is to audit tracking setup before running a study, especially for offline conversions, GA4 imports, and app conversions.

The Conversion Lift documentation, updated on November 13, 2026, did not change the incrementality test's methodology, but it tightened the technical tracking requirements that feed into the result. The update itself now explicitly lists unsupported configurations, and Google's own recommendation is to audit tracking setup before running a study, especially for offline conversions, GA4 imports, and app conversions.

What Google Ads changed in the Conversion Lift guide

Google updated, on November 13, 2026, its help documentation on how to set up Conversion Lift studies. The information comes from Search Engine Roundtable, which compared the current version of the "Set up Conversion Lift based on users" document with the page's historical archive and published a diff of the changes.

Conversion Lift is the test Google Ads offers to measure incrementality: it splits the audience into an exposed group and a control (holdout) group to estimate how much of the conversion happened only because of the ad, something last-click attribution doesn't answer. It's the tool growth teams use when the question isn't "how many conversions came from this campaign," but "how many additional conversions did this campaign generate."

The update doesn't touch the test's statistical design. Study duration, minimum budget, conversion thresholds, control group size, accepted bidding strategies, and campaign selection criteria remain the same as in the previous version. What changed was the data layer that feeds that design: Google detailed, with much more precision, what counts as valid tracking.

The six technical points that can invalidate a study

The documentation reinforces requirements on six fronts. Each one is a place where a setup that seems to work may, in practice, be delivering insufficient data for the lift calculation to be reliable.

• Enhanced conversions: more detailed requirements on hashed customer data (email, phone, physical address) and on correctly sending transaction IDs.
• Offline conversions: import based solely on GCLID is no longer sufficient; the guidance now is to use Enhanced Conversions for Leads with hashed first-party data.
• Unsupported setups: the document now explicitly names configurations incompatible with Conversion Lift, including web conversions imported via GA4.
• Consent Mode: more specific guidance on Advanced Consent Mode and measurement eligibility in EEA, UK, and DMA scenarios.
• App conversions: more detailed requirements for attribution integrations on Android, iOS, Firebase, and third-party tools.
• Data quality: it's recommended to have 7 to 14 days of continuous conversion data before starting the study, send dynamic conversion values with valid currencies, and route between 75% and 80% of web conversion traffic through the Google tag gateway.

In this article's reading, the items about GA4 and GCLID tend to be the ones that most catch legacy accounts off guard: operations that still rely on GA4 conversion import or attribution based solely on GCLID may have built that pipeline before Enhanced Conversions became established as common market practice.

Why this is an ROI problem, not just a technical checklist

For those who decide channel budget, the consequence is direct: a Conversion Lift study run on a setup that Google itself now classifies as unsupported can report incrementality that doesn't exist, or underestimate the incrementality that does exist. The decision to scale or cut a campaign, made on top of that number, inherits the instrumentation error.

In summary: the update isn't saying "Conversion Lift changed its methodology," it's saying "part of the tracking that feeds the test is probably incomplete, and that can distort the result." The update's own source sums up the main message: the growth team should audit the tracking setup before launching a study, especially if it relies on CRM/offline conversions, GA4 imports, or app conversions. In practice, this means treating that audit as a prerequisite, not an optional step.

The documentation point that's easiest to audit is the Google tag gateway one: the recommendation to route between 75% and 80% of web conversion traffic through it is a number any martech team can check directly in the tag panel, without relying on interpretation. It's this kind of objective metric that separates a real audit from an assumption that "tracking is fine."

What stays the same (and why that makes fixing it easier)

Since the study's statistical design hasn't changed, anyone who has already run Conversion Lift before doesn't need to relearn the methodology or redesign the experiment from scratch. The work now is instrumentation auditing, not a new testing process. This lowers the cost of compliance: it's one more tracking checklist before hitting the button to launch the study, not a change to the growth process.

Practical checklist before launching or relaunching a study

Before starting a Conversion Lift study, it's worth auditing, in this order:

1. Confirm whether the lead or offline conversion uses Enhanced Conversions, and not just GCLID.
2. Check whether any relevant conversion comes from a GA4 import; if it does, it's outside Conversion Lift support.
3. Check the Consent Mode configuration, especially for traffic from the EEA, UK, or markets under the DMA.
4. For apps, review the attribution integration (Firebase, native SDK, or third-party) against the per-platform requirements.
5. Ensure at least 7 to 14 days of continuous, stable conversion data before the study starts.
6. Measure, in the tag panel, whether the web conversion traffic routed through the Google tag gateway is between 75% and 80%.

What the growth team should take from this

For those who report incrementality to stakeholders, the update is a reminder that the lift number is never more reliable than the data pipeline that feeds it. It's worth treating this audit as routine, not as a one-time task: any migration from GA4 to another conversion source, change of app attribution vendor, or change in consent rules can reopen one of the six points listed by Google.

Anyone with an active Conversion Lift study who hasn't reviewed the instrumentation since before Enhanced Conversions became common practice has a concrete reason to run this checklist before the next budget cycle, because it's the incrementality data, not the campaign itself, that's at stake.