Connected TV vendors have every commercial reason to show advertisers near-perfect completion rates - and near-perfect rates are exactly what the format delivers by default, regardless of whether anyone watched.

Austin Lake, a paid media director and consultant who describes managing more than $100 million in ad spend across Google Ads, Meta, programmatic, and CTV, raised the point directly in a LinkedIn discussion thread this month. His position is not new in ad tech circles, but the framing is pointed: on connected television, a 98% completion rate is not a sign of campaign success. It is the structural baseline of the medium.

Why the number is almost always 98% or above

The mechanics are worth understanding. Unlike pre-roll video on mobile or desktop browsers, the dominant ad formats on streaming platforms are non-skippable. A viewer watching a FAST channel or a subscription-supported streaming service typically cannot dismiss an ad once it starts. The ad plays. The impression is logged. The completion is recorded. The system reports success.

This is why Vlad Chubakov, Associate Director of Programmatic at Delve Deeper, noted in November 2025 that completion rates on streaming environments "almost always reach 99.99%." His conclusion: "Some marketers use it to show they are meeting KPIs, but it is meaningless for optimizing performance or quality."

The FreeWheel benchmark cited by the Video Advertising Bureau, drawn from September 2023 research, put the average ad completion rate for long-form premium video at 94%. That figure, high as it is, reflects content environments where viewers are more engaged and less likely to leave. In open-market programmatic CTV, where inventory quality varies widely, the number may look similar on paper while telling a very different story about who, if anyone, was present.

The vendor reporting problem

Dr. Augustine Fou of FouAnalytics put the conflict of interest plainly in the same LinkedIn thread. Advertisers should never trust the performance reports provided by the CTV vendor they are buying from, he wrote. "What else would they say other than the campaigns are performing great, keep spending."

This is not a theoretical concern. PPC Land reported in March 2026 that DoubleVerify data suggested more than one in three CTV impressions deliver to a screen that is switched off. The HDMI-CEC protocol - built into virtually every HDMI-connected device since 1999 - is technically capable of signalling whether a television is powered on. In practice, the advertising measurement stack almost never queries it. The ad-serving layer, the SSP, the DSP, and the verification vendor each operate on data from the streaming device, not from the screen itself. A 100% completion rate on a CTV impression can mean a human watched an ad from start to finish. It can equally mean a set-top box executed a delivery log entry while the television sat dark.

DoubleVerify, in its May 2026 Global Insights report, found that CTV fraud schemes and variants rose 140% in Q1 2026 compared with Q1 2025. The company identified more than 50 distinct bot attacks and variants in 2025 alone. More than that: 34 out of every 100 CTV impressions DoubleVerify monitored did not appear in streaming TV content at all, instead serving on music apps, fitness apps, screensavers, and teleconferencing platforms - environments that carry streaming device identifiers but bear no resemblance to what advertisers believe they are buying. Only 21% of surveyed advertisers currently use invalid traffic or fraud detection as a key performance indicator for CTV campaigns.

What completion rate actually incentivises

The commercial logic of optimising on completion rate compounds the problem. Because the metric is structurally guaranteed on non-skippable formats, algorithms that chase it will gravitate toward the cheapest available inventory that produces a logged view. That inventory is not premium streaming. It is the open programmatic market: FAST channels with high supply volumes, mixed content quality, and limited publisher transparency.

Chubakov's November 2025 analysis identified this as one of four critical errors in CTV campaign execution. Cost per Completed View, which functions as a proxy for CPM, and completion rate together create an incentive structure that rewards the lowest-cost impression rather than the most valuable one. The IAB Europe guide to programmatic CTV, published in April 2026, made the same argument from a measurement standards perspective: CPM alone is not a reliable proxy for value in connected television, and more meaningful indicators include incremental reach, frequency distribution, completion-rate stability, attention-weighted exposure, and outcome-based proxies.

Austin Lake's reply to Dusty Sutherland in the thread also illustrates a practical wrinkle. Sutherland noted that newer placements - specifically the FireTV feature rotator - can bring down completion rates below the near-universal baseline. Lake's response: those placements should go on their own line item, similar to pause ads. The point is not only methodological. If completion rates vary by placement type, then a blended campaign-level completion rate obscures rather than reveals where the delivery actually happened and what environment it appeared in.

SMBs and enterprise alike are flying blind

Lake added a separate observation in the thread, responding to Fou's comment about vendor self-reporting. Most marketing teams at small and mid-sized businesses, and even at some enterprise companies, are not aware that independent measurement options exist. The default posture - trusting the reports supplied by the vendor selling the inventory - is the norm, not the exception.

This structural gap has a scale implication. eMarketer projects CTV ad spending to reach $38 billion in 2026, representing 43% of all television advertising budgets. The Video Advertising Bureau, in its June 2026 multiplatform report, documented that 209.4 million US viewers now use ad-supported streaming services. That is a channel operating at significant scale, with a measurement framework that has not kept pace.

The IAB Spain study from May 2026 found that 44% of advertising professionals expressed neither much nor little confidence in current CTV measurement systems, with 8% reporting no confidence at all. The two most commonly used KPIs were reach (66%) and view-through or completion rate (65%). Purchase intent and store visit lift appeared in only 13% and 7% of professional survey responses respectively. That distribution describes an industry still measuring what is easy to count rather than what is useful to know.

What the industry is building instead

The recognition that completion rate is insufficient has driven considerable product development. Teads launched deterministic CTV measurement in October 2025, enabling advertisers to track site visits, leads, and sales directly tied to CTV exposure - moving attribution beyond impressions and completed views for the first time outside the United States. OpenX and TVision launched pre-bid attention targeting for CTV in March 2026, repositioning attention data from a post-campaign analysis tool to a real-time buying signal. Index Exchange embedded a comparable pre-bid attention signal in February 2026 through a separate partnership with xpln.ai.

The IAB's October 2025 guide on Conversion APIs urged industry-wide adoption of server-to-server data frameworks to connect CTV exposure to actual business outcomes - purchases, leads, app installs - rather than logging the delivery and calling it done. The IAB Europe CTV conversion gap analysis from March 2026 identified the absence of universal identifier standards and cross-device attribution infrastructure as the primary blockers preventing CTV from operating as a measurable performance channel.

None of that infrastructure eliminates the baseline problem Lake and Fou identified. A vendor report showing a 98% completion rate remains, in most circumstances, a description of the ad format rather than a description of campaign effectiveness. The rate tells the advertiser that the non-skippable unit played to its end. It says nothing about whether the television was on, whether anyone was in the room, whether the creative registered, or whether any downstream action followed.

Timeline

  • September 2023 - FreeWheel research documents a 94% average ad completion rate for long-form premium video content
  • November 4, 2025 - Vlad Chubakov publishes LinkedIn post identifying completion rate and CPCV as misleading KPIs for CTV; Dr. Augustine Fou engages in comments
  • October 23, 2025 - Teads launches deterministic CTV measurement enabling outcome tracking beyond impressions and completion rates
  • October 30, 2025 - IAB releases 26-page guide urging standardised Conversion API adoption for CTV
  • February 2026 - Index Exchange embeds pre-bid attention signal into its SSP via xpln.ai partnership
  • March 11, 2026 - OpenX and TVision launch pre-bid CTV attention targeting across 231 million monthly unique users
  • March 14, 2026 - IAB Europe publishes CTV conversion gap analysis identifying identifier fragmentation as primary measurement blocker
  • Late March 2026 - Michael M. Maurantonio raises HDMI-CEC screen-state detection as a CTV measurement problem; Dr. Augustine Fou responds
  • April 2026 - IAB Europe publishes Guide to Programmatic for CTV, arguing CPM alone is not a reliable value proxy
  • May 6, 2026 - IAB Spain publishes fifth annual CTV study showing 44% of professionals lack confidence in CTV measurement; completion rate remains the second most common KPI at 65%
  • May 7, 2026 - DoubleVerify publishes 2026 Global Insights report documenting 140% rise in CTV fraud schemes in Q1 2026 and 34 in 100 impressions not landing in streaming TV content
  • June 2026 - VAB documents 209.4 million US viewers on ad-supported streaming; eMarketer projects CTV ad spend at $38 billion for 2026
  • June 2026 - Austin Lake raises CTV completion rate baseline problem in LinkedIn discussion; Dr. Augustine Fou and Dusty Sutherland contribute comments

Summary

Who: Austin Lake, a paid media director and consultant reporting more than $100 million in managed spend, with responses from Dr. Augustine Fou of FouAnalytics and Dusty Sutherland, an advertising strategist.

What: A LinkedIn discussion this month challenged the standard use of completion rate as a CTV key performance indicator, arguing that non-skippable ad formats structurally guarantee rates near 98% or above regardless of viewer attention, inventory quality, or actual campaign effectiveness. Vendor-supplied reports have a built-in commercial bias toward showing these numbers favourably.

When: The LinkedIn thread appeared in June 2026, against a backdrop of documented CTV fraud increases (140% in Q1 2026 per DoubleVerify), persistent industry confidence gaps in CTV measurement (44% of professionals in IAB Spain's May 2026 study), and ongoing infrastructure development aimed at replacing delivery metrics with outcome measurement.

Where: The discussion originated on LinkedIn and applies across the connected television advertising ecosystem, including FAST channels, premium streaming environments, programmatic open market buys, and direct platform deals.

Why: With CTV ad spending projected by eMarketer at $38 billion in 2026 and completion rate remaining the second most commonly used CTV KPI among advertising professionals, the distinction between a structurally guaranteed metric and an earned one has direct implications for how budgets are allocated, which inventory is favoured, and whether advertisers can identify the difference between a completed impression and a watched ad.