LiveRamp and the Marketing + Media Alliance (MMA) published joint research on July 20, 2026. It shows that small identity errors can make a profitable B2B campaign look like a loss. In one test, a campaign with a true $1.50 return measured as just $0.43. That drop is steep enough to get a working campaign cancelled.
What LiveRamp and MMA Found in The Missing Piece Report
The report is called “The Missing Piece: Improving Confidence in Marketing Measurement.” It used real impression logs and synthetic data to test two common measurement problems. The first is missing data: ad exposures that happened but never got recorded. The second is identity mismatch: exposures recorded but linked to the wrong person.
Random data loss turned out to be mostly harmless. Dropping 20% of impressions at random left channel rankings unchanged. Non-random loss did real damage. That is the kind caused by privacy controls or a broken tracking pixel. It flipped channel rankings at just 1% of total volume. The model’s own accuracy score even improved while its answer got worse. A marketer watching the dashboard would see the numbers look better as the conclusion became false.
The identity findings carry the sharper warning. Researchers ran a simulation where half of all matched users were linked to the wrong person. A true 25% lift measured as roughly 6.8%. A true $1.50 return measured as $0.43. LiveRamp and MMA call this an illustrative lower bound, not a typical result. Still, it shows how the error can quietly wreck a measurement read. Randomized tests that count everyone assigned to a group, whether reached or not, held up best against both problems.
Where the LiveRamp and MMA Report Comes From
LiveRamp (NYSE: RAMP) is the San Francisco data collaboration company behind the study. It ran the research with MMA, the nonprofit alliance representing more than 825 marketing organizations. Business Wire distributed the full announcement on July 20, 2026. Both companies presented the findings that same day at the MMA CMO + CEO Summit in Santa Barbara. MMA’s SVP of Research and Insights, Vassilis Bakopoulos, said the research points to two blind spots. Missing data hides which channels really work. Identity matching, he said, “feels like progress but isn’t built for precision.”
Why Measurement Trust Is Already Under Pressure
The report lands at a moment when B2B marketers already doubt their own numbers. A recent industry study found that only 15% of marketers use effectiveness data to set budgets. Many say they don’t trust what the data tells them, a gap covered in our report on the WFA and Ebiquity effectiveness study. Separately, LinkedIn scored 700 B2B campaigns on seven buyability signals. That research found the inputs feeding a score matter as much as the creative itself. And Anteriad’s 2026 report found that the marketers pulling ahead are the ones who control their data and buying-group signals. Identity errors sit underneath all three problems. Bad matching corrupts the very data marketers are being told to trust more.
Talking Shift: Bad Data Kills Good Campaigns Before Anyone Notices
A 70% swing in measured ROI is not a rounding error. It is the difference between a campaign getting more budget and a campaign getting cancelled. Start Some Shift’s view: marketers spend enormous effort proving a campaign worked, then trust the same measurement system without checking whether it matched the right people in the first place. “You cannot out-strategize bad identity data. If the number is wrong, the decision built on it is wrong too,” says Start Some Shift. Before marketers ask whether a campaign worked, they need to ask whether the system that measured it actually knows who saw the ad.
Practical Implications for B2B Marketers
Marketing leaders should treat this report as a prompt to audit measurement before trusting it with budget decisions.
- Reconcile publisher-reported delivery against your own measurement logs, especially in regional campaigns where audience files skew by ZIP code.
- Ask every measurement vendor how they verify identity precision, not just how large their match rate is.
- Stress-test your measurement systems for non-random data loss, since most audits only check random loss.
- Run the odds ratio check the report recommends, comparing conversion odds for exposed versus unexposed users.
- Favor randomized, intent-to-treat test designs for high-stakes budget calls, since they held up best in the study.
- Flag any campaign recommended for cancellation based on one measurement read, and re-test before pulling spend.
- Build identity precision requirements into vendor contracts, not just data volume or match-rate targets.
What to Watch Next
Watch for reaction from measurement vendors and holding companies. LiveRamp is itself the subject of a pending $2.5 billion Publicis acquisition, and this report recommends its own product category as the fix. Expect more marketers to start asking measurement partners for identity precision numbers, not just match rates, over the next two quarters.
