Google’s new Data Strength Uplift metric reports Google’s own calculation of the conversions your data setup recovered, and Google’s announcement of it lists causal proof as a separate step. That distinction is the useful part of the 10 September 2026 update to the Google Ads measurement suite. The metric is a report from the platform that sells the ads. Proof that your ads caused a sale still comes from a test.
None of that makes the new tools a bad idea. It does mean the averages Google published alongside them should be read with their footnotes, and that a small business paying for Google Ads should keep one count of results that Google does not produce.
Want this built for you?
Animus Engine installs the marketing brain and the engine that runs on it. A human hand on every gate.
What Google Announced on 10 September 2026
Google announced a set of measurement changes in one post, from new data connections to a new metric inside Google Ads. Nipoon Malhotra, Google’s vice president for ads analytics, insights and measurement, wrote the announcement on the Google blog.
- Data Manager, Google’s tool for connecting customer data, is being integrated into Google Analytics and Display & Video 360.
- Enhanced conversions are launching in Google Analytics and Display & Video 360.
- The Data Manager API, the programmatic route for sending your own customer data to Google’s ad products, is now universal, based on an industry standard from the IAB Tech Lab.
- A new Data Strength Uplift metric in Google Ads “calculates the additional conversions recovered by your first-party data setup”.
- Meridian, Google’s open-source marketing mix model, gets agentic features and brand signals, and its GeoX library for causal geo-experiments is now generally available.
The post frames all of this as moving measurement “from a reactive report card to a proactive performance engine”, resting on three elements: a strong data foundation, multiple signals, and causal proof.
Tell Us the Task That Eats Your Week
Name one marketing job you’d hand off tomorrow. We’ll send back how we’d run it with AI — the steps, the tools, and where a human stays in charge. A real reply from Chris within one business day.
The Four Averages in Google’s Announcement, and What They Measure
Every figure in the post is an average from Google’s own data, and each carries a footnote with a different time window. Read together, they tell you what Google observed across advertisers who made a change. They do not tell you what the same change will do in your account.
The Four Figures in Google’s 10 September 2026 Measurement Post, With Their Footnoted Limits
| Figure | What Google says it measures | Whose data, and which period | What it is not |
|---|---|---|---|
| 26% | Average increase in incremental return on ad spend, the revenue the ads added per unit of spend beyond sales that would have happened anyway, for advertisers who connect offline and app data to Data Manager | Google data, global, April 2025 to April 2026, Search campaigns bidding to conversion value | A result for any single account, or for campaigns not bidding to conversion value |
| 11% | Average increase in Search conversions for enhanced conversions compared with standard conversion imports | Google internal data, global, 1 to 14 January 2026 | A long-run figure: the window is two weeks |
| 14% | Average conversion uplift for advertisers who build data strength with Google tag gateway | Google internal data, global, July to December 2024 compared with January to June 2025 | A controlled test: it compares two different half-years |
| Over 20% | Uplift for Demand Gen campaigns, Google’s campaign type for image and video ads on YouTube, Discover and Gmail, among those advertisers | Google data, global, 3 to 17 June 2026 | A Search campaign figure, or a long-run one |
The footnotes name a data source and a period for each figure, and none of them describes a randomized comparison. The last column is our reading of the footnotes, not Google’s wording. A figure measured over two weeks, or across two half-years, can move for reasons that have nothing to do with the tool.
A Report Card Is Not Causal Proof
Google’s own framework separates the report from the proof: the Data Strength Uplift metric calculates recovered conversions, while causal proof comes from experiments such as the geo-experiments in Meridian GeoX. The announcement says so in its structure. Data foundation and signals sit in one part of the post, and “causal proof” arrives with GeoX, which Google describes as its library “for running causal geo-experiments across any advertising platform”.
In plain terms, a recovered conversion is one the setup lets Google count that it would otherwise have missed. That can be true and still be a different question from whether the ad caused the sale. The customer who bought after clicking might have bought anyway. Only a comparison against a group that did not see the ads answers that, which is why Google lists it as a separate element.
Consider an owner who switches on enhanced conversions and watches the conversions column climb week after week, while the invoice book shows the same month as last year. Nothing in that account is broken. Google is now counting sales it used to miss, and that owner has to decide at budget time whether better counting is worth more spend.
What a Small Business Can Check About Google Ads Data Strength Itself
Four steps keep your own count beside Google’s, and none of them needs new software. The platform can count and estimate, but someone who knows the business decides whether a change worked.
- Record a full month of conversions and sales in your own records before you change your data setup.
- Change one part of how data reaches Google Ads at a time, so a change in results has one likely cause.
- Wait a full comparable period before reading results.
- Compare the conversions column against your own sales, note the Data Strength Uplift figure beside it as Google’s estimate, then decide.
The waiting step is the one worth protecting. Two of Google’s own figures come from windows of about two weeks, and a short window is exactly where a normal swing in sales looks like an effect.
What to Do With the Next Measurement Launch
Read the next measurement launch the way this one reads best: with its footnotes open and your own sales records beside it. We looked at the wider question of whether the channel pays for a small business in is Google Ads worth it for a small business.
Next step: write down last month’s conversions and sales side by side, then book a strategy call and bring both numbers.