Google's Data Strength Metric Counts Recovered Conversions

Google returned to conversion recovery in its October 13 Demand Gen update. For a small SaaS team reviewing its AI campaign budget, that creates a reporting question. Does the dashboard show a better count of existing leads or additional customers?

The October update highlights Data Strength Uplift, a Google Ads metric for the conversion gains associated with an advertiser’s data setup. It repeats a claim of more than 20% average conversion uplift among Demand Gen advertisers using Google Tag Gateway, based on Google’s internal global analysis covering June 3-17, 2026.

These features were already announced on September 10. October’s post is a follow-up, and it gives no new controlled evidence that the claimed reporting gain means additional sales. The earlier announcement defines the metric as conversions recovered through the advertiser’s first-party data setup.

Stephen Chang, Google’s first-party data management and measurement specialist, linked more customer signals to better AI performance in September. They said:

“the better Google AI can perform.”

That explains why a marketing team would improve its data connection. The budget decision still needs evidence about what the ads caused.

The sale can precede its appearance in the report

Google Tag Gateway changes how the measurement code reaches a visitor’s browser. In the same conversation, Chang explains that the tag can be served through the advertiser’s own domain, instead of a Google-hosted domain. The stated benefit is recovering conversions the platform would otherwise miss.

Data Manager handles a related job: bringing customer data into Google’s advertising products, including records that weren’t captured by the website tag. For a lead-generation business, that can mean connecting later sales milestones to the original online inquiry. Google uses those records for measurement and for campaign optimization.

Natalia Bohm, Google’s conversion solutions specialist, described advertisers without enhanced conversions as:

“missing out on critical conversions”

Google’s enhanced-conversions documentation describes the matching mechanism. Advertisers send customer information, such as an email address, after a one-way transformation called hashing. Website data can be matched to signed-in Google accounts that interacted with ads. Imported offline lead conversions can be linked back through customer data collected on the website.

A reporting system with more complete records gives the advertiser a better account of outcomes. A recovered record can also give automated bidding another useful example of the customer it’s meant to find. Neither fact establishes how many purchases would disappear if the campaign stopped running.

The percentage needs its own boundary

The more-than-20% figure is Google’s average across the Demand Gen advertisers in its cited analysis. The public footnote supplies a global scope and a two-week window. It supplies no sample size, randomized control, or method for separating additional sales from additional reporting.

Google sells the advertising and the measurement tools involved. Its specialists explain the product’s intended benefits; their statements aren’t independent validation of the performance claim. The number is a reason to investigate conversion capture in an account, with the account’s own baseline, rather than a forecast to paste into a spending proposal.

There’s also an availability boundary. October’s update says Data Manager is expanding into Google Analytics and Display & Video 360. It describes enhanced conversions in those products as forthcoming. An announcement isn’t proof that every advertiser has the same controls available today.

These distinctions belong in the report shared with finance. Record when the data connection changed, which conversion action changed, and whether the outcome means a submitted lead or a closed sale. Comparing totals across that implementation date without an annotation makes a measurement change easy to mistake for a campaign improvement. We found a related comparability problem in YouTube’s modeled view counts.

Better capture has a separate test from better advertising

The strongest case for Google’s update is practical: incomplete records make it harder to bid toward a valuable outcome. For a SaaS team, connecting qualified leads and later sales to the campaign gives bidding a more useful objective than form submissions alone.

Google also documents a way to ask the causal question. Conversion Lift compares people exposed to ads with a control group that isn’t exposed. The difference estimates outcomes driven by the ads. It supports user-based and geographic study types, although the documentation says the tool isn’t available to every account and directs advertisers to their Google representative.

A team without access should keep that limitation visible. A stronger data connection and a larger conversion total don’t substitute for the unavailable experiment. That principle also applies to the AI measurement evidence discussed in our Ahrefs coverage: understand how a number was collected before assigning work or budget to it.

For the next marketing budget review, annotate the capture change and check the outcome feeding automated bidding. Keep the recovered-conversion result in the measurement assessment; require separate evidence about additional qualified customers before increasing spend on that basis.