The Video Advertising Bureau closed its 2026 summer school series on August 28 with an email to marketers headlining a 94% average ad completion rate for long-form premium video, a figure drawn from FreeWheel research published in September 2023.

The message went out from VAB Insights at 14:30 on Friday, August 28, 2026, under the subject line "Today's Lesson: Impressions." It was the final instalment of a seasonal email series aimed at brand marketers and agency planners. The topic was impressions, and specifically the argument that impressions delivered on premium video platforms carry more value for advertisers than impressions delivered elsewhere.

One number carried the argument. According to VAB, the average ad completion rate for long-form premium video content is 94%, a figure the trade body presented as evidence of how engaged audiences are when advertisements run inside premium environments. The email linked to four reports from the organisation's library: The Power of Premium Video, The Impression Gap, The Illusions of the Internet, and You Oughta Know.

Where the 94% comes from

The completion figure is not new research. It appears on page 34 of The Impression Gap, the custom study VAB published earlier this year, and the slide carries its own source line: FreeWheel, The Delicate Art of Balancing Ad Load, September 2023.

That places the number roughly three years before the email that promoted it. FreeWheel, the ad-serving business owned by Comcast, has been measuring completion rates on premium video inventory for longer than that. A FreeWheel Signature Insights report covered by PPC Land in early 2022 put the average completion rate across connected television at 97%, a figure the company at the time described as a sign of high engagement across connected devices.

The gap between those two readings is worth noting. A 97% figure for connected television generally and a 94% figure for long-form premium video specifically are not directly comparable, but neither reading supports a claim that premium environments produce unusually complete ad views relative to the rest of the channel.

What a completion rate measures

Completion rate counts whether an advertisement played to its end. On most connected television inventory, the format does not permit anything else.

That structural point was the subject of a PPC Land report on June 29, 2026, which examined a public discussion among media buyers about whether completion rate functions as a performance metric at all. The argument set out there was that non-skippable formats guarantee rates near 98% regardless of inventory quality, viewer attention, or whether anyone is in the room. A completion rate describes the ad unit. It does not describe the audience.

Standards bodies have moved in a similar direction on the adjacent question of attention. The Media Rating Council and the Interactive Advertising Bureau published attention measurement guidelines in November 2025 establishing minimum requirements for quality, transparency and comparability across vendors. Those guidelines stated explicitly that attention should not be treated as a measure of campaign outcomes, describing it instead as an exposure and engagement signal.

Completion sits a further step removed from outcomes than attention does. It records delivery.

The study underneath the email

The Impression Gap itself rests on original measurement rather than licensed statistics. VAB commissioned TVision to supply custom research comparing how audiences consume connected television content on premium video platforms against YouTube. PPC Land reported the findings when the study was released on February 24, 2026, and examined the platform-level detail again in June.

The methodology covers 21 premium video platforms spanning hybrid AVOD/SVOD services, virtual multichannel video programming distributors, and free ad-supported streaming television. The fieldwork ran from July 1, 2024 to June 30, 2025 across viewers aged two and over. TVision collected the data through in-home sensors performing person and facial recognition, automatic content recognition fingerprinting to identify what was on screen, and a digital meter detecting which device and application were in use.

Three index metrics structure the analysis, each demanding more of the viewer than the last. The presence to active index measures the share of time a viewer was in the room out of the total active session. The attention to presence index measures the share of time spent with eyes on screen while the viewer was in the room. The attention to duration index measures the share of time eyes stayed on screen relative to the length of the session, with a floor of 30 seconds of consistent viewing.

Averaged across the 21 platforms, premium video scored 102 on presence to active against 97 for YouTube, 108 against 95 on attention to presence, and 110 against 93 on attention to duration. Co-viewing, defined as the proportion of impressions occurring with two or more viewers holding overlapping active sessions for at least five minutes, reached 60% on premium platforms against 45% on YouTube. Average session length was one hour and 18 minutes against 52 minutes.

The session threshold is asymmetric

A detail in the study definitions table deserves attention from anyone reading the session length comparison. For a session to be included in the analysis at all, it required a minimum duration of 30 seconds on YouTube and five minutes on premium video platforms.

That is a tenfold difference in the inclusion floor. Sessions between 30 seconds and five minutes enter YouTube's average and are excluded from the premium video average. Since those sessions are by definition the shortest, the asymmetry pulls YouTube's reported average down and lifts the premium figure. The 49% session length advantage and the 26-minute absolute gap both sit on top of that construction.

The same five-minute overlap requirement applies to the co-viewing metric, which is measured on a platform whose median session is materially shorter.

Averages conceal per-platform rank

VAB's summary characterises premium video as outperforming YouTube across every metric measured. That holds at the level of the category average. It does not hold platform by platform, and the study's own charts show where it breaks.

On the presence to active index, YouTube ranked 14th of the 19 charted entries at 96.8, placing it above five individual premium platforms scoring between 94.5 and 96.1. On attention to duration, YouTube again ranked 14th at 92.6, above five premium entries running down to 65.2. On attention to presence, YouTube ranked 17th at 95.0, above two premium platforms. On co-viewing, YouTube ranked 18th at 45%, above one premium platform at 42%.

Only on session length did YouTube rank last outright.

The charts show 19 entries rather than 21 because four Spanish-language applications were grouped into a single analytical unit. On the attention to duration chart, that grouped Spanish-language entry scored 65.2, the lowest reading in the dataset by a wide margin and 27 points below YouTube.

None of this contradicts the direction of the averages. It does mean that a planner choosing between a specific streaming application and YouTube cannot read the category result as a guide to that particular trade.

A dating inconsistency in the source material

The deck's own source notes are not internally consistent on the measurement window. Every slide except one dates the custom study to July 2024 through June 2025, and the study definitions table specifies July 1, 2024 to June 30, 2025. The footnote beneath the co-viewing chart on page nine instead reads June 2024 to July 2025, a 13-month window offset by one month at each end.

Two further slides carry source notes referring to an "Attention to Visible Index" indexed to a TV viewing norm, while the metric guide and every other slide define an attention to presence index indexed to a connected television viewing norm. VAB's landing page for the report adds a third variant, stating that the study uncovers which platforms work hardest for marketers on connected television, "which are now present in 83% of US TV households" - a construction in which the antecedent of the relative clause is ambiguous.

The supporting statistics are borrowed and dated

The completion rate is not the only third-party figure in the deck, and the pattern it illustrates runs through the supporting evidence.

According to the report, Magnite research from November 2025 found that 43% of consumers say they are likely to discuss advertisements with others when co-viewing. Thinkbox UK research from 2024, conducted with Map The Territory and Tapestry Research across 4,005 viewing occasions among 2,017 respondents aged 18 to 75, recorded a 23% increase in ad recall when watching with others in the living room rather than alone. Dentsu's What is the Attention Economy, published in January 2024, is cited for a claim that attention metrics improve return on investment campaign forecasting by 38% compared with viewability. WARC's July 2024 attention guide, drawing on a Mars Essence Mediacom study, supplies a 30% higher sales lift and 10% return on investment increase for attention-optimised media. Adelaide's 2025 Outcomes Guide from January 2025 supplies a 41% higher brand lift figure.

The most recent of those six external citations dates to January 2025. The oldest, the completion rate, dates to September 2023. The original TVision measurement is the newest material in the document, and it closed on June 30, 2025 - fourteen months before the email that promoted it.

Named voices in the report

The deck carries quotations from five industry figures, each attached to a definitional slide rather than to the underlying data.

Vikrant Mathur, co-founder of Future Today, is quoted on co-viewing: "Co-viewing leads to deeper engagement with ads and, as a result, higher performance...ultimately, when family members are in a room watching together, it increases their attention, not just for programming, but for advertising, as well."

Jason Jutla, head of practice for EMEA and the United Kingdom at WPP Media, addresses planning: "Attention gives us a unifier across the different channels that we operate in...which will help us from the planning phases all the way through to our optimization on an ongoing basis."

Mike Follett, chief executive of Lumen Research, states the case for the metric category: "What people look at is strongly linked to what they buy. Attention is the missing link between exposure and effectiveness."

Karen Nelson-Field, chief executive of Amplified Intelligence, addresses duration: "The more seconds of active attention an ad receives, the more days it can stay in the memory, and the longer it can work."

James Rooke, president of Comcast Advertising, supplies the line closest to the completion argument: "Ads viewed in the long-form, lean-back TV environment have greater unaided recall and purchase intent."

Rooke's employer owns FreeWheel, the source of the 94% figure.

Why this matters for media buyers

Connected television is now a large enough budget line for benchmark quality to have consequences. eMarketer projections cited in VAB's March 2026 streaming analysis put United States connected television advertising spend at 38 billion dollars in 2026, equivalent to 43% of all television advertising budgets, rising to a projected 47.6 billion dollars and 53% by 2028.

Buyer confidence has not kept pace with that spend. Research published by the IAB in July 2026 found that 43% of connected television buyers report only somewhat to no confidence in knowing where their advertisements ran, even in the formats they trust most, with the figure climbing above 60% once private marketplaces and open exchanges are included.

That uncertainty is precisely why delivery metrics retain their appeal. A completion rate is easy to report, arrives in every log file, and reads as high. It is also, on non-skippable inventory, close to a constant.

The measurement market has been moving away from it. Attention signals have migrated from post-campaign reporting into bidding, with pre-bid attention targeting reaching connected television supply during 2026 and attention scores entering impression-level bid valuation. TVision itself changed hands in April 2026, acquired by Viant for 40 million dollars, a transaction PPC Land has covered as part of a broader shift away from self-graded television measurement.

The interest behind the argument

VAB represents premium multiscreen television providers and distributors. Its members sell the inventory the research favours, and the report is distributed to members and what the organisation describes as qualified marketers. That does not make the TVision measurement wrong - the sensor data is collected by a third party and the index construction is documented in the deck. It does mean the framing around the numbers is advocacy rather than neutral analysis, and the selection of which external statistics to include is part of that framing.

The three-year-old completion rate is the clearest instance. It is the number the email chose to lead with, it is the only headline statistic in the message, and it measures the thing least connected to the argument the rest of the document makes about attention.

Timeline

Summary

Who: The Video Advertising Bureau, a New York trade organisation representing premium multiscreen television providers and distributors, distributing to marketers, agencies and its own members. The underlying measurement came from TVision, now owned by Viant. The completion figure originated with FreeWheel, the Comcast-owned ad server.

What: An email headlining a 94% average ad completion rate for long-form premium video content, presented as evidence that impressions on premium video platforms outperform impressions elsewhere. The figure dates to FreeWheel research from September 2023 and appears inside The Impression Gap, the VAB and TVision study measuring 21 premium platforms against YouTube on co-viewing, three attention indices and session length.

When: The email was distributed at 14:30 on August 28, 2026. The custom study fieldwork ran from July 1, 2024 to June 30, 2025, and the study itself was published on February 24, 2026.

Where: The United States connected television market, covering hybrid AVOD and SVOD services, virtual MVPDs and free ad-supported streaming platforms measured through in-home TVision sensors.

Why: Connected television spend is projected at 38 billion dollars in the United States in 2026, or 43% of all television advertising budgets, while 43% of buyers report limited confidence in where their advertisements ran. Delivery metrics such as completion rate remain widely reported because they are easy to produce, even as standards bodies and buy-side practitioners argue that non-skippable formats make near-total completion a property of the format rather than a measure of audience engagement.