CAPIs Won't Solve CTV's Outcomes Problem Until Another Key Issue Is Fixed
Netflix's recent conversion API announcement was met with a predictable chorus of approval. After all, the industry wants to see streamers stepping up to prove outcomes. To many, it's a clear signal that TV is finally growing up and speaking the language of performance marketing.
But let's not get ahead of ourselves with the self-congratulatory back-slapping, as there's a more pressing question that we should be asking. Before we measure what an ad achieved, shouldn't we first be able to confirm what the ad actually ran against?
Advertisers need more than assurances about media quality
CTV media buying, in 2026, still largely operates on trust. Buyers make their decisions based on promises about audience composition and content environment, assessing brand suitability and making direct deals. The trouble is that these deals rarely come with the verification to back them up.
Ask most advertisers whether their last CTV campaign ran alongside the shows and genres they planned for, and they won't be able to tell you. Genre-level reporting and show title verification are baseline expectations in digital display environments, but in CTV, they remain largely aspirational. That's a media quality problem, and CAPIs don't even come close to solving it.
The performance conversation in streaming TV has become almost entirely focused on the back end. Advertisers want conversion tracking and attribution modelling so that they can prove that an impression led to an outcome. And while this work is important, it's being built on a foundation that hasn't been properly inspected.
Without placement verification, CAPIs are of little worth
If you can't verify that your ad ran in the context you bought it in, then what you're really measuring is the outcome of a partially unknown media buy. Your measurement system tells you what happened downstream, but it can't tell you whether the targeting that was supposed to shape that outcome was ever actually in place.
Viewers in different content environments are in different mindsets, and as such, their levels of receptivity to different types of advertising will differ too. A financial services ad running in a news environment will reach a viewer in a distinct frame of mind compared to one running mid-way through a reality show. An automotive brand buying premium drama placements doesn't expect their ad to be shown alongside broad-reach entertainment programming.
And the performance metrics downstream will reflect these misalignments. Without placement verification, advertisers can't learn from their buying decisions and they certainly can't optimise against them. What's more, they can't hold supply partners accountable for poor performance caused by ads being shown outside of their intended context.
Why CTV's metadata problem persists
The reasons for this issue are no great mystery; the dominance of direct deals in CTV means there is less call for standardised, machine-readable metadata that more programmatic-led channels would demand. Fragmented supply across hundreds of apps and services means content classification is inconsistent even when it exists. Publishers might provide general signals like genre or language, but often stop short of disclosing show-level data.
The culture of upfront buying has compounded this. Relationships and reach have historically mattered more than post-campaign verification, so the industry has found it easy to kick the transparency question into the long grass. This deferral made some sense when CTV buys were lumped together with traditional TV buys, when limited reporting was the norm. But it makes much less sense now.
The industry is now inviting performance-oriented advertisers into a buying environment that wasn't built for scrutiny. Digital-native brands arriving at CTV via self-serve platforms like those offered by Roku and Comcast don't have the same tolerance for opacity that traditional TV buyers developed over decades; they will want to know what they bought. These advertisers will expect to be able to verify their buys, which is a problem because the infrastructure to give them that answer at scale doesn't exist at present.
Media quality and measurement are interdependent
There may be a temptation to treat media quality verification and outcome measurement as sequential problems to solve; once the plumbing is fixed, we can worry about what flows through it. But the reality is that these aspects are interdependent. Better contextual transparency means better targeting verification. Better targeting verification means better-informed optimisation. Better-informed optimisation means better outcomes. We don't have to wait for attribution to improve before we start seeing the benefits of cleaner supply-side data.
What's needed is a commitment from publishers and platforms to treat show and genre-level reporting as a standard post-campaign deliverable in CTV, not an optional add-on. The technical capability exists; the IAB's OpenRTB protocol already supports content object signals that can carry programme-level data. The same industry that built viewability and brand safety standards for digital display is capable of building equivalent standards here.
CAPIs are not the villain here; the performance instinct they represent is exactly what CTV needs to develop as a serious channel for outcome-driven advertisers. But performance requires a clean foundation. Confirming what was actually bought, in which context, against which content, is a prerequisite for the kind of rigorous, accountable advertising that the performance conversation assumes is already in place.
[Editor's note: This is a contributed article from Adlook. Streaming Media accepts vendor bylines based solely on their value to our readers.]