Winterberry Group on September 30, 2026 published The State of Unified Marketing Measurement, a survey of 121 US marketing executives at vice-president level and above, sponsored by Lifesight. Eighty-three percent of respondents rated cross-channel spend optimization a leading priority, according to Winterberry Group, while only 9% said they could practice unified marketing measurement "extremely well."

In Short

A research firm asked 121 senior US marketers how well they can work out which of their advertising channels really drive sales once all the channels are looked at together. Most called it a leading priority, but only about one in 11 said they do it extremely well, and almost all of them still rely on reports made by the same platforms that sell them the ads. The study was paid for by Lifesight, a company that sells exactly the kind of independent measurement the report says brands need, so its numbers come with that context attached.

A small, senior sample

The fieldwork ran in July and August 2026. According to Winterberry Group, the proprietary online survey reached 121 US marketers and media professionals at the VP level and above who hold responsibility for cross-channel marketing measurement. They worked in financial services, insurance, multichannel retail, direct-to-consumer brands, consumer packaged goods, travel and hospitality, automotive, cable and telecommunications, and consumer technology. Annual marketing budgets in the sample ranged from $5 million to more than $250 million. Results were aggregated for presentation, and the anonymous quotes scattered through the report are attributed by job title and industry only.

Lifesight commissioned the work as part of what the two companies describe as a broader research effort into how brands measure, attribute and optimize performance across channels. The sponsorship is not incidental. The 16-page document's pages carry a Lifesight copyright line, the report is distributed from Lifesight's website, and its seventh chapter is given over to a description of Lifesight's own product. Winterberry Group, a New York-based management consultancy, conducted the survey.

No margin of error is published. For a simple random sample of 121, the conventional 95% margin on a result near 50% is roughly plus or minus 9 percentage points. The report does not describe how respondents were recruited, and that calculation assumes random selection, so it understates rather than overstates the true uncertainty. Small percentages also translate into small headcounts: 9% of 121 is about 11 executives, and 16% is about 19.

The headline figure and its footnote

The central contrast in the companies' announcement - 83% against 9% - rewards a closer look. A footnote explains that the 83% figure counts respondents who scored 6 or 7 on a seven-point scale running from "not at all a priority" to "a leading priority." The question asked how far their organization prioritizes optimizing the allocation of advertising and marketing spend across paid and owned channels. It does not use the phrase unified marketing measurement. The announcement from the two companies treats that answer as equivalent to identifying UMM as a leading priority.

The wording behind the 9% figure is not published at all. Readers therefore cannot tell whether the two numbers come from the same question, the same scale, or the same respondents.

A separate question produces a lower but arguably sharper number. When respondents were asked to name their top measurement priorities, 40% chose unified measurement to optimize cross-channel spend, according to the report. That put it ahead of higher-quality data sources at 33% and better incrementality and causality measurement at 31%. Forty percent naming it first is a different statement from 83% rating it highly; both are true within the survey's design.

The report's own glossary defines unified marketing measurement as an approach that integrates multiple methodologies and data sources to produce a holistic, cross-channel view of performance that guides investment decisions. The announcement narrows that to a specific trio: causal marketing mix modeling, geo-based incrementality testing and incrementality-adjusted attribution. Those are the same three components Lifesight says its platform combines.

Where the gap widens

The most technically useful table in the report compares mean priority and mean ability scores, both on a 1 to 7 scale, across five capabilities.

CapabilityPriorityAbilityGap
Optimizing spend allocation across paid and owned channels6.15.2-0.9
Understanding causality behind marketing outcomes6.05.3-0.7
Optimizing content and offers for audience segments5.64.9-0.7
Measuring the relative impact of multiple channels5.85.2-0.6
Measuring the impact of individual channels5.65.3-0.3

The pattern is consistent. The narrowest gap sits at the single-channel level, where platform reporting tools do most of the work; the widest sits at cross-channel allocation, the decision about where the next dollar goes. An insurance EVP quoted in the report described an organization with "strong single-channel measurement" but "almost no credible view of how channels interact across the full funnel." A retail CXO put it from the operator's side: "The single biggest challenge is getting everything that we need from one superior platform, instead of having to use multiple platforms to reach our marketing goals."

The report draws a distinction between consolidation and unification. Pulling every number into one dashboard, it argues, still leaves conflicting numbers side by side; unification means the methods are calibrated so that they agree.

Five methods and no referee

Last-click attribution, once the industry default, is now used by only 32% of enterprise brands in the sample. Its place has been taken by a crowded toolkit. Platform-native reporting is used by 93%, marketing mix modeling by 71%, multi-touch attribution by 69%, data-driven attribution by 62%, incrementality testing and panel-based research by 52% each, and custom methodologies by 43%. Summed, those shares come to 4.74 methods per brand, which the report rounds to "nearly five."

Brands are broadly content with the parts. They gave their current measurement an average score of 7.0 out of 10 for how well it supports business goals, according to the report, and 64% rated their measurement capability at 6 or higher on the same 10-point scale. Sixty-eight percent said measurement insights now guide more than half of their marketing spend, and 73% said insights drive all or a substantial part of their marketing decisions.

"Brands aren't starting from scratch. Most have invested in sophisticated measurement capabilities and are already using those insights to inform their marketing decisions," said Jonathan Margulies, managing partner at Winterberry Group. "The opportunity now is to move from measuring individual channels to optimizing how those channels work together as part of a seamless media mix. That shift, from better measurement to better optimization, is where we expect to see brands focus most of their investment and effort in the years ahead."

How do these adoption rates compare with other studies? Not neatly. When PPC Land covered Meta's suite of truth white paper in May 2025, the Kantar research cited in it put the average advertiser at 3.8 measurement solutions, with 55% experiencing contradictory results across them. The MMM figure diverges more sharply. PPC Land's report on Amazon's MMM API reaching general availability in 14 markets in May 2026 cited research putting MMM implementation at just 15%, with only 8% of in-house teams holding the advanced analytics skills required. Winterberry's 71% comes from a senior, enterprise-only sample, which may explain much of the distance, but the spread is a reminder that using MMM can mean anything from a quarterly consultancy deliverable to an always-on in-house model.

The walled-garden arithmetic

The report opens with Maya, a composite VP of marketing at a national retailer, facing a CFO who asks where an extra $5 million would go and how anyone would know it worked. Her team has pulled numbers from Meta Ads Manager, Google and a retail media network. The accompanying chart shows Meta claiming 41% of conversions, Google 38% and retail media 34% - a combined 113% of actual sales. More sales than the retailer actually made! The chart is an illustration built around a composite character, not survey data, and the report presents no underlying measurement for it.

The survey numbers that follow are real enough. Ninety-three percent of brands use platform-native reporting, and 84% of those with dedicated measurement technology call it a core part of their stack. More than three in four said they were satisfied with their ability to measure individual channels, scoring themselves 5.3 out of 7. Around that core, 82% use data warehouses such as Snowflake or BigQuery, 66% use BI tools, 65% run custom-built models in Python or R, and nearly half use clean rooms or MTA platforms.

The structural backdrop is concentration. According to the report, roughly 59% of trackable US advertising spend flowed directly to Alphabet, Amazon and Meta in 2025, up from 47% in 2020. The report does not cite a source, though the figures match those in PPC Land's August 2025 analysis showing the Amazon, Google and Meta triopoly taking 58.8% of total US ad spending, against 47.1% five years earlier. The executive summary separately puts US media investment at nearly $700 billion in 2026, also without a cited source.

An SVP in consumer technology summarized the concern: "Walled garden platforms increasingly function as both the media seller and the sole source of truth." Tobin Thomas, chief executive of Lifesight, put it more bluntly in the announcement. "Every platform grades its own homework, and marketers are left reconciling numbers that were never meant to reconcile," he said. "Unified measurement means one independent, causal view that marketing and finance can both plan against."

The report then adds figures that do not come from the survey at all. In Lifesight's work with more than 300 brands, it states, platforms routinely overcount their contribution by 20% to 50%, and around 40% of spend often turns out to be non-incremental once measured causally. No methodology, time period or client breakdown accompanies those claims. They are vendor estimates, presented alongside the survey results without distinction. This is where the report brings in iROAS - incremental return on ad spend, defined in its glossary as the revenue a channel caused divided by what was spent on it - as the corrective to platform-reported ROAS.

Independent research has pointed in the same direction, with harder evidence. PPC Land reported on September 26 that an analysis of 663 Facebook experiments found machine-learning lift estimates reaching a median of 83% for upper-funnel outcomes, against 29% in randomized tests. But the problem is not confined to platforms. A Zalando researcher's paper, covered in August, showed a conventional MMM reporting 10.61x return on paid search against a true value of 4.20x, a 2.5-fold overstatement that geo-experiment data corrected. Marketing mix models, in other words, can inherit the same bias they are meant to correct.

The platforms have also moved into building the models themselves. PPC Land's April analysis of Meta's open-source Robyn documented a case in which adopting the tool coincided with a 15% to 20% lift in Facebook's share of one advertiser's budget. On September 9, Google moved Meridian GeoX out of beta, claiming budget savings of more than 31% for large advertisers on geo experiments. In November 2025, Google had already cut the minimum budget for incrementality experiments from roughly $100,000 to $5,000. Causal measurement is becoming cheaper to access - and more of it is being supplied by the companies whose media it evaluates.

One number, two framings

The announcement and the report present the 93% figure differently. The announcement links it to the staffing shortfall: "To fill this gap, 93% of these marketers are relying on measurement tools built and hosted by media platforms to assess individual channels or campaigns." Read literally, that describes 93% of the 74% who need outside help, choosing platform tools as a substitute for expertise. The report presents 93% as a base rate across the whole sample, with no causal link to staffing. Nothing in the published data supports the announcement's "to fill this gap" framing.

AI: widespread, rarely central

Seventy-nine percent of brands use AI to support measurement, according to the report. The leading uses are generating insights and summarizing results at 85%, forecasting and scenario planning at 77%, and data preparation at 74%. Only 16% said AI plays a central role in their measurement.

The report describes 48% as taking an AI-assisted, human-led approach and calls that "most," though it is a plurality rather than a majority. Another 25% describe their approach as purely AI-led. That figure sits awkwardly beside the 16% who say AI is central; the report does not reconcile the two. Recommending budget and media actions, and executing them, remain predominantly human tasks.

The barriers are mostly about inputs rather than models. Poor or inconsistent data quality was cited by 55% as an obstacle to acting on AI-generated measurement insights, limited integration with decision and transaction systems by 50%, and lack of trust in results by 40%. Further down came the cost of AI utilization (37%), limited resources to oversee AI tests (36%) and methodologies that cannot accommodate AI insights (30%). A financial services CXO offered the most succinct summary: "Integrating AI recommendations into media planning lacks a formal process and needs structure."

The report's argument here is worth stating plainly. An AI agent summarizing platform-reported ROAS will summarize numbers that overcount; one recommending budget shifts from a stale, correlation-based model will produce correlation-based recommendations faster. The constraint, it contends, is the evidence underneath.

That echoes earlier work. The IAB's February study of more than 400 senior decision-makers found up to 75% of marketers saying attribution, incrementality and MMM underperform, and estimated AI could unlock about $32 billion in measurement value. In August, the Coalition for Innovative Media Measurement published a 52-page report on how MMM embedded in automated budget systems can amplify data errors at scale, proposing six governance steps.

The people constraint

Dependence on measurement raises the cost of falling behind. Only 18% of brands said they have all the people, skills and expertise they need to meet their measurement objectives, according to Winterberry Group. Seventy-four percent said they need outside help to expand capacity and capability. The two most common reasons were interpreting results and translating them into budget or media recommendations, cited by 69%, and executing methodologies such as MMM and geo experiments, cited by 64%.

"For us, technology is not the constraint here. It's talent, good measurement people are difficult to find and harder to retain," an EVP in retail told the researchers.

The finding is consistent with earlier industry surveys. TransUnion and EMARKETER reported in October 2025 that 60.2% of organizations faced internal stakeholder skepticism about measurement, with 28.6% saying it put 11% to 20% of budgets at risk. The CIMM and 4As study of 197 senior marketers at companies with budgets above $50 million, covered by PPC Land in March, found 43% citing cross-platform measurement gaps as a major obstacle. Across these studies the shortfall looks less like missing data than missing capacity to act on it.

What the sponsor sells

The report's sixth chapter sets out three moves for closing the gap: building a culture that acts on insight rather than data, choosing partners who stand outside the walled gardens, and aligning marketing, finance, analytics and operations on what is being measured and why. The second move calls for independent partners with no media-buying affiliation who can measure the relative, incremental and complementary contribution of every channel. That is the description Lifesight applies to itself.

The seventh chapter makes the link explicit. It maps each finding to a requirement - nearly five unconnected approaches to calibrated causal MMM, platform dependence to independent measurement, the 69% seeking interpretation help to "recommendations expressed as decisions, not dashboards" - before stating: "This is the standard Lifesight was built to meet." The report closes with an invitation to book a demo.

Lifesight describes itself as an Agentic Unified Marketing Measurement Platform combining causal marketing mix modeling, geo-based incrementality testing and incrementality-adjusted attribution. Its Marketing Intelligence Agents, according to the company, interpret results, forecast outcomes, recommend reallocations and push approved changes back to ad platforms with one click. A feature called Ask Lifesight accepts questions in plain English. The company says it is independent of any media-buying platform and compliant with GDPR, HIPAA, SOC 2 and ISO 27001.

Its performance claims are presented without supporting detail. According to Lifesight, more than 300 global brands use the platform, with an average sales lift of roughly 9.8%; one client, Seidensticker, generated 11.5% more revenue on 11.7% less spend after reallocating on causal evidence. No time frame, baseline or method is given for either figure.

The client count also differs from what the company said four months earlier. When Lifesight released a read-only connector built on MCP in June, PPC Land reported that Lifesight cited more than 500 global brands and over $4 billion in combined marketing spend, alongside a claim of at least a 20% improvement in return on ad spend. The current materials say more than 300. Neither the report nor the announcement explains the difference. In August, PPC Land's review of a Lifesight report on CTV and paid search found that its headline 22.3% conversion-rate lift came without methodology, sample sizes or confidence measures.

None of this invalidates the survey. The fieldwork was run by a third party, the sample is described, and several findings are reported with useful precision. But the report blends three kinds of material: survey data from 121 executives, Lifesight's own client estimates, and a composite narrative. The epilogue returns to Maya a quarter later, her causal model having identified retail media as capturing sales that would have happened anyway and a geo-lift test having confirmed the model's reading on CTV. It is a tidy ending. It is also a composite scenario, flagged as such only in the prologue.

Why the marketing community is watching this

For media buyers, the practical question behind the report is who arbitrates between conflicting numbers when budgets move. Platforms are offering more causal tools at lower thresholds - Google's Meridian and GeoX, Meta's Robyn, Amazon's MMM API - while independent vendors argue that a seller cannot neutrally evaluate its own inventory. The Winterberry data puts a figure on how far enterprise marketers say they are from resolving that tension.

The disputes are already channel-specific. In April, the IAB and Instacart argued in a white paper, citing February research, that commerce media is underrepresented in half of MMM implementations and that closed-loop incrementality measurement deserves priority for retail media budgets. The report's epilogue, in which a causal model shows retail media capturing sales that would have happened anyway, points the opposite way. Which methodology a brand trusts can decide which channel wins the next budget cycle - and every party offering a methodology also has a view on the answer.

Timeline

Summary

Who: Winterberry Group, a New York-based management consultancy, conducted the research; Lifesight, a marketing measurement vendor, commissioned and sponsored it. Jonathan Margulies, managing partner at Winterberry Group, and Tobin Thomas, chief executive of Lifesight, provided statements. Respondents were 121 US marketing and media executives at VP level and above.

What: The State of Unified Marketing Measurement found that 83% of respondents rate cross-channel spend optimization a leading priority while 9% say they practice unified measurement extremely well. Ninety-three percent use platform-native reporting, 79% use AI in measurement but 16% call it central, 18% have all the expertise they need and 74% want outside help. The report mixes survey data with Lifesight's own unverified client estimates and a composite narrative.

When: The survey was fielded in July and August 2026, and the report was published on September 30, 2026.

Where: The survey covered US enterprise brands across nine industries with annual marketing budgets from $5 million to more than $250 million. The report is distributed through Lifesight's website.

Why: Marketers face conflicting numbers from platforms that both sell media and measure it, while platform-built tools such as Meridian, Robyn and Amazon's MMM API expand. The report quantifies the gap between how much enterprise marketers prioritize cross-channel measurement and how well they say they can do it - a gap its sponsor sells a product to close.