Connected TV was supposed to fix what traditional television never could: opaque measurement, broad assumptions, and probabilistic outcomes built on GRPs. And in many ways, it has. But the real question is not whether CTV is more measurable than linear TV — it clearly is. The more meaningful debate is whether CTV measurement has truly reached the sophistication long associated with digital advertising. As brands demand clearer incremental impact, stronger signals, and full-funnel accountability, CTV is both advancing measurement methodologies and revealing structural weaknesses in the ecosystem. Dan Larkman, CEO and Founder @ Keynes, offers his take on the matter.
CTV has long promised better measurement than traditional TV. Do you believe CTV measurement has truly caught up to the sophistication of digital advertising?
Firstly, CTV offers significantly more measurement than traditional TV. Traditional TV was built on GRPs and broad assumptions. If you served four ads to a household, you assumed a transaction would follow. The signals were basic and probabilistic at best.
When the comparison shifts to digital, the conversation changes. I would argue that CTV actually has the opportunity to make digital more sophisticated. Many brands built their measurement models on click attribution within a single device and treated it as definitive proof of performance. Click has been used as the be-all and end-all metric, even though it was never truly sophisticated. It is an easy metric, but not necessarily a needle-moving or incremental one.
What CTV has accomplished over the past two to three years is to push the industry toward more advanced measurement. The growth of geo-lift studies, synthetic control groups, MMM frameworks, and other advanced methodologies has been significant. Those models are advancing as brands seek to demonstrate the value of all channels and the incremental value of each channel.
The challenge is that CTV measurement has multiple links in the chain. With a platform like Meta, the journey can be one device, one site, one transaction. In CTV, there are more variables. And when any link in that chain is weak, the system breaks down quickly.
Self-serve CTV tools have created click-like proxies and rely heavily on inherently inaccurate IP addresses. Recent CIMM data suggest IP accuracy ranges from 13 to 16 percent. That creates structural fragility. Measurement can be highly sophisticated, but if the foundation relies on weak identifiers, the signal falls apart.
So the real issue is not whether measurement can be sophisticated. It is whether the underlying signals are strong enough to support it.
How do streaming platforms’ measurement frameworks compare to those of walled gardens like Meta, Amazon, or TikTok? Are they stronger or weaker?
Some streaming platforms are stronger, and the distinction becomes clear when you look at the data being surfaced. It varies by provider, but certain CTV platforms present order IDs, which Meta does not. That level of transparency matters.
Providing order IDs allows advertisers to distinguish new-to-file customers from existing customers, understand the exposure-to-conversion window, and evaluate how many touches are required before conversion. That depth of insight creates a more accountable measurement framework.
The walled gardens tend to operate with more of a “trust me” mentality. They are highly protective of their audiences because that audience data is their core asset. They structure their measurement frameworks to prevent reverse engineering of that audience. Meta, for example, does not share order IDs. As a 100 percent mobile platform, it has also had to navigate significant data privacy constraints, particularly following Apple’s iOS 14.5 changes. That shift forced the development of predicted view conversions and other modeled approaches, which inherently rely on assumptions.
There is also a financial incentive behind that lack of transparency. For example, providing deeper visibility could reveal that a platform should account for 35 percent of the marketing budget rather than 50 percent. That is a meaningful shift, so there is limited incentive to expose that level of detail.
Amazon presents a similar dynamic. Their audiences command significant premiums compared to standard audience pricing. They may perform well, but the question is whether they are cost-effective relative to their performance.
In the CTV ecosystem, there is often more data available and more that can be surfaced. Given the fragmented, highly competitive landscape, streaming platforms must demonstrate value more directly. They cannot rely solely on brand dominance or institutional comfort. That creates the opportunity for greater transparency and a more robust measurement framework.
Walled gardens can lean on scale and the reality that no one gets fired for running a Meta or Amazon campaign. Streaming platforms, by contrast, must earn that trust through rigorous measurement.
As brand budgets continue shifting into CTV, what must evolve in measurement to prove true sales impact, and what does the future look like in 2026?
It is less about shifting brand budgets and more about evaluating the entire marketing funnel as a single system. Historically, branding teams managed television, radio, and out-of-home, while performance teams focused on clickable platforms like Meta. Those silos are starting to break down.
Branding and performance teams are now working more closely together because CTV supports both objectives. Creative can drive performance or build brand, while still carrying the equity of television. That overlap creates opportunity but also tension.
The teams operate on different metrics. A branding team might look at metrics such as reach and frequency or brand lift studies. A performance team may expect the same return on ad spend as Meta. The value of CTV often shows up in the long tail rather than in immediate, direct-response metrics, and measurement must reflect that.
Time to conversion becomes critical. Understanding how long it takes from first exposure to transaction, and reducing that window, directly impacts customers’ lifetime value. For brands with multiple purchase cycles per year, shortening that path materially accelerates growth.
At the same time, the ecosystem remains heavily reliant on IP addresses as a cross-device signal. As long as IP remains the dominant identifier, measurement will stay imperfect. It is deeply embedded, so replacing it will not be simple, but meaningful progress requires moving beyond IP as the primary foundation.
The future of CTV measurement is about aligning brand and performance accountability and strengthening the signals that underpin full-funnel impact.
