A measurement report published by Lifesight with four connected TV sellers argues that streaming ads are moving revenue through channels that never get the credit. The document, circulated to press on August 12, 2026, puts the paid search conversion lift from paired CTV exposure at 22.3% and the paid social lift at 8.5%, and it leans on anonymized case studies rather than disclosed methodology.

Lifesight has published "The State of CTV Measurement 2026: Navigating the Full-Funnel of CTV," a fifteen-page report developed with four companies that sell connected television advertising: Roku, Universal Ads, Vibe.co and MNTN. The document was distributed to trade press on Wednesday, August 12, 2026. Its argument is narrow and consistent throughout: the constraint on connected television budgets is not the channel's effectiveness but the measurement apparatus pointed at it.

The central figure in the report concerns spillover. According to Lifesight, campaigns paired with CTV produce a 22.3% conversion rate lift in paid search and an 8.5% increase in paid social conversions. Those numbers appear twice in the document, once in a highlighted callout and once inside a section on causal marketing mix modeling. No sample size, time period, vertical composition, or statistical method accompanies them. Neither does a source citation.

That absence sets the pattern for the report as a whole, and it is worth stating plainly at the outset. This is a vendor document. Lifesight sells unified marketing measurement software. Roku, Universal Ads, Vibe.co and MNTN sell connected television inventory or the tooling used to buy it. Every party named on the cover has a commercial interest in the conclusion that CTV works better than existing dashboards suggest.

The no-click argument, quantified

The report's framing rests on a structural asymmetry. A social ad produces a click. A television ad, on a screen with no cursor and no browser, does not. Attribution systems built to trace clicks therefore under-record a channel that generates no clicks by design.

To quantify what conventional reporting misses, the document assembles five engagement metrics. Logged-in CTV sessions average 99 minutes, which the report describes as 12 times the length of an average eight-minute social media session. Ad completion on Paramount CTV inventory reaches 96%. Research using EEG data, the report states, shows brain alertness among CTV viewers at 95.3% against 56.2% for social media viewers, close to a two-fold difference. Nine in ten CTV viewers are described as being in a good mood while watching, while one in four social media users report a negative mood while scrolling. And 76% of users seek new content on streaming television compared with 47% on social.

One of these figures does not hold still across the document. Page two puts the 96% completion rate against 16.4% for standard video ads and 30% to 40% for Instagram Reels. Page five puts the same 96% against 53% "typically seen on social media." The comparison base shifts by a factor of three between pages, which matters because the headline ratio depends entirely on which denominator is used. The report does not reconcile the two.

Sourcing is thin throughout. The EEG study is unnamed. The mood research is unattributed. The 96% Paramount figure carries no methodology note. For a document whose thesis is that marketers accept numbers too easily from platform-reported dashboards, the evidentiary standard applied to its own statistics is notably lighter than the standard it recommends for everything else.

Credit versus cause

The conceptual distinction the report presses hardest is between attribution and causality. Attribution assigns credit for a conversion that touched a given channel. Incrementality asks whether the conversion would have happened without the exposure. According to the report, most teams still evaluate CTV inside the ad platform using platform-reported sessions, platform-reported return on ad spend, and platform-reported conversions.

The document offers a specific illustration of the resulting confusion: a platform such as Vibe.co and a tool such as GA4 will often report different session counts, because they measure different aspects of viewer behavior. The report argues that the discrepancy itself is not the problem, and that the problem is drawing a firm conclusion from a single data source without validating it through testing.

Jacob Sailer, Technical Account Manager at Vibe.co, framed the same point in the report: "The brands that get the most out of CTV stop asking 'Did CTV get credit?' and start asking 'Did CTV move the number?' Those are very different questions. Only one of them tells you what to do next."

Dan Lapinski, Director of Ads Manager Product Commercialization at Roku, tied the shift to budget defense. "CTV has earned its place on the plan, now advertisers need proof it's earning its share of the budget," he said in the report. "At Roku Ads Manager, we're focused on enabling causal measurement - incrementality testing, geo-lift studies, and closed-loop commerce signals - because the brands making bigger CTV commitments are the ones who can tie exposure directly to business outcomes, not just impressions to clicks."

That positioning is consistent with Roku's product direction over the past two years. Roku Ads Manager arrived in September 2024 as a self-service platform aimed at performance marketers accustomed to search and social buying. In January 2026, Roku became the first major streaming publisher to optimize campaigns on iSpot outcomes rather than reach metrics, with SimpliSafe test data showing a 23% lift in leads. By June 2026, Smartly had connected to Roku Ads Manager through the Roku Ads API, the platform's first API-linked demand partner.

Andy Everson, VP of Global Partnerships at MNTN, addressed the reliability of incrementality methods themselves: "Attribution tells you where a conversion touched; incrementality tells you whether it would have happened anyway, and closing that gap depends largely on the quality of the partners behind the measurement." He added that methods built for the display era do not transfer cleanly to streaming, where exposure is harder to observe and clean control groups are harder to hold.

Four case studies, unevenly documented

The report's evidence base consists of four campaign examples, each disclosed at a different level of detail.

The first is anonymized. A brand entered an incrementality study through Vibe.co convinced that CTV would over-index on net-new customer acquisition. The test measured a 3.03x incremental ROAS, but the majority of that incremental revenue came from existing customers rather than new buyers. The team initially read the result as a disappointment. According to the report, that was the wrong conclusion, because the channel was generating exceptional incremental revenue from an audience segment that already knew the brand, at a return that outperformed several other channels. No advertiser name, category, spend level, test duration, or confidence interval is given.

The second involves wellness brand Obvi, which ran a geo-lift test to determine whether Universal Ads activity was driving in-store sales at Walmart. Markets were matched for similarity and split into exposed and control groups. The report states that Obvi measured an incremental lift in in-store purchases but does not quantify it. Ashvin Melwani, CMO and Co-Founder of Obvi, is quoted saying: "For the first time, we could see the real impact of CTV on retail sales, not just what we hoped it was driving."

The third case is the most numerically specific. A California-based real estate firm running TV, radio, Meta, Google and direct mail spent roughly $350,000 on television in a single quarter without knowing whether that spend correlated to revenue. A marketing mix model incorporating saturation curves, ad stock effects and mediator variables identified a point of diminishing returns: leads had stopped growing while costs continued to rise. According to the report, the resulting reallocation cut acquisition costs by 45% while scaling leads by 14%.

The fourth concerns Image Skincare, which layered premium content bundles on top of existing MNTN activity. After the layering, the brand's MNTN incremental ROAS rose 26%. The report presents this as evidence that upper-funnel exposure made lower-funnel conversion work harder rather than competing with it for the same budget.

Three of the four studies name the advertiser. None discloses methodology, control design, or measurement window. Each was produced by or with a company that appears on the report's cover.

The second screen as a measurement problem

A section of the report deals with what it calls second-screen attribution. Nearly two-thirds of US social network users now watch television or streaming while simultaneously scrolling on their phones, according to the document. Within that behavior, 31% of viewers have used the second screen to shop via QR codes or links during a show, and 34% have purchased merchandise, collectibles or fashion inspired by a movie or TV show.

The operational consequence is that program selection carries commercial weight beyond the ad slot itself, since content influence can drive purchases independently of the advertisement. The report also notes that AI-based visual search is being built directly into flagship LG and Samsung sets, allowing viewers to turn a television frame into a scannable interaction.

The industry has been building rails for this behavior for some time. QR code overlays in non-skippable YouTube ads on connected TV arrived through Display and Video 360's Instant Reserve function in May 2024. In June 2026, Samsung Ads made in-stream Samsung TV Plus ads addable to an Amazon cart with a TV remote. At Brandcast on May 13, 2026, YouTube went further, introducing two-click checkout on the television itself via Google Pay, explicitly to remove the second-screen hop. The measurement question the Lifesight report raises therefore has a shrinking window in some environments and a widening one in others.

Spend growth and the CPM discrepancy

On market size, the report states that CTV ad spend has increased 207% since 2020, that 71% of marketers raised CTV budgets in 2025, and that a 2026 survey suggests 70% of CTV advertisers will continue increasing spend by an average of 17%. It also states that CTV ad costs have fallen since 2023 to roughly $25 to $35 CPM, with premium placements and networks commanding $50 to $65.

Those pricing figures sit well above what programmatic auction data shows. DataBeat figures reported in early 2026 put average programmatic CTV CPMs at $4.82 in January 2026, up 29.2% year over year and higher than every other device category. The gap is a definitional one rather than a contradiction: managed-service and direct-sold television inventory prices differently from open-auction programmatic supply, and the report does not specify which market its range describes.

Recent earnings data adds a further complication to the "costs are falling" framing. Roku's second quarter 2026 results, published on August 6, 2026, showed 40% more video ad impressions sold at 12% lower unit prices, a volume-for-price trade indicating that supply growth continues to outrun demand. MNTN, meanwhile, reported second quarter revenue of $82.5 million on August 4, 2026, growth of 21% against an IAB projection of 13.8% for the connected television category in 2026.

Why the timing matters for buyers

Three of the four contributing companies are inside corporate transitions that affect how durable any measurement commitment made through them is likely to be. Universal Ads sits in an unresolved ownership position following Comcast's June 29, 2026 announcement of a tax-free spin-off separating NBCUniversal and Sky from its connectivity business; the filing did not specify which resulting company retains the platform. Vibe.co agreed to be acquired by Walmart on June 23, 2026, a deal covering more than 10,000 advertisers and subject to Hart-Scott-Rodino clearance. Roku is the subject of an acquisition by Fox Corporation, which is why the company held no earnings call for the second quarter.

The measurement argument the report makes is not new, and the pressure behind it is well documented. IAB research covered in July 2026 found that 43% of CTV buyers have only limited confidence in where their ads actually ran. Jamloop opened a household-level holdout methodology on July 21, 2026, reporting 3,224 incremental subscriptions at 99.98% confidence for an unnamed streaming service. Nexxen bundled incrementality into its DSP in March 2026, Innovid added purchase-impact and control-group measurement in April 2026, and Smartly signed a letter of intent to acquire INCRMNTAL in March 2026 to embed always-on incrementality into its buying stack.

Lifesight itself is not new to this territory either. The company appeared among twelve vendors profiled in IAB Australia's market mix modeling landscape study in September 2025, and on June 2, 2026 it launched an MCP connector giving Claude and ChatGPT access to live measurement models for customers managing more than $4 billion in combined marketing spend.

Martha Ann Pavoni, VP of Product and Measurement at Universal Ads, argued in the report that the lift percentage is the less interesting half of an incrementality result: "The value of incrementality measurement isn't just the lift percentage. It's revealing who you reached that you otherwise wouldn't have. In many of our client studies, a meaningful share of incremental conversions come from customers the brand wasn't already reaching."

What the report does not contain is any disclosure of how its own headline figures were produced. The 22.3% and 8.5% conversion lifts are the numbers most likely to be quoted from it in budget conversations, and they are the numbers with the least documentation behind them. For an argument built on the premise that platform-reported figures deserve validation through testing, that is a conspicuous gap.

Timeline

Summary

Who: Lifesight, a marketing measurement software company, produced the report with four connected television companies: Roku, Universal Ads, Vibe.co and MNTN. Named contributors include Dan Lapinski of Roku, Jacob Sailer of Vibe.co, Martha Ann Pavoni of Universal Ads, Andy Everson of MNTN, and Ashvin Melwani of the wellness brand Obvi.

What: A fifteen-page report titled "The State of CTV Measurement 2026: Navigating the Full-Funnel of CTV," arguing that click-based attribution systematically under-records connected television's contribution to revenue. Headline claims include a 22.3% paid search conversion lift and an 8.5% paid social lift when campaigns are paired with CTV, a 3.03x incremental ROAS in an anonymized test, a 45% acquisition cost reduction alongside 14% lead growth at a California real estate firm, and a 26% incremental ROAS increase at Image Skincare. None of the headline statistics carries a source citation, sample size, or methodology note, and the ad completion comparison base differs between pages two and five.

When: The report was circulated to trade press on Wednesday, August 12, 2026. It carries a 2026 Lifesight copyright.

Where: The data described is primarily United States market data. Case studies reference Walmart retail locations, a California-based real estate firm, and campaigns run through Roku, Universal Ads, Vibe.co and MNTN platforms.

Why: Connected television budgets have grown to a size where finance teams demand proof of causation rather than correlation, and the channel produces no clicks to trace. The report is positioned into that gap, though every organization named on its cover sells into the channel it evaluates, and three of the four contributing platforms are currently inside pending acquisitions or corporate separations.