Connected TV offers countless benefits to marketers. Ad targeting is made simpler since advertisers have access to data about users’ category preferences and the shows they watch. This enables them to serve relevant ads to target audiences based on this information. And, of course, viewers have been binging streaming content, so there’s no question that the audience is there.
But in the new high-definition world of content, ad measurement hasn’t yet come into full focus. With so many players in the mix – including connected devices from makers such as Roku, Apple TV and Chromecast, and an ever-expanding pool of streaming services with their own operating systems – it’s easy to see why measurement is no easy feat.
The challenges of accurate CTV ad measurement
The challenges of ad measurement in streaming stem primarily from the fragmentation of the ecosystem. Put another way: there’s no common currency.
While Nielsen’s framework worked seamlessly within the well-oiled machine of linear TV, there are no such accepted standards of measurement in CTV. Traditional TV audience measurement methodologies, like Nielsen’s panel-based approach for linear TV, were designed for large audiences watching a relatively small universe of channels. In a sprawling, addressable, and highly dynamic CTV ecosystem, that approach doesn’t give advertisers the insight they need into campaign performance, say experts.
A key issue is the lack of common identifiers across the different CTV platforms. This further impedes standardization. CTV media buyers are clamoring for a universal, cross-platform identifier to make planning and measuring CTV campaigns easier.
Still, the central problem is that most streaming services have their own proprietary methods of measurement. And on top of these approaches are countless third-party vendors offering ad verification and measurement services. As a result, it’s increasingly difficult for advertisers to gain a clear and comprehensive picture of ad performance or to track metrics such as attribution.
“For advertisers and content creators using the established advertising-based video on-demand (AVOD) social platforms such as YouTube and Facebook, the measurement challenge is more about confidence in the data and access given to third-party measurement,” says Graham Swallow, head of data and insights at digital content agency Little Dot Studios. “We don’t know what YouTube defines as a ‘view’, and we don’t have any independent verification of the view data, nor even the simple ability to look at cross-platform reach or frequency.”
For video on-demand (VOD) companies – whether they are device manufacturers like Samsung and Vizio or streaming platforms like Roku – there is a similar lack of standardization. “Data collection is often a manual process, with each platform providing different levels of granularity and richness, without agreed definitions of metrics,” Swallow says. “Advertisers and agencies want ease and confidence that campaigns will deliver and numbers to back them up. There are opportunities in this space to solve these challenges.”
There are the proprietary approaches to measurement by the ‘big four’ (Roku, Amazon, Apple and Google), which contribute to the challenges advertisers face in painting a holistic picture of streaming ad performance. Taken to another level, buyers can transact across hundreds of different apps, aggregators and publishers, further splintering audiences and complicating measurement.
The fragmentation of the market makes it difficult to measure with accuracy. Tal Chalozin, the chief technology officer and co-founder of adtech company Innovid, says: “It’s impossible to really understand who the viewer is who’s actually watching – and if it’s the same viewer who is watching on two different devices or two apps on the same device, versus scores on multiple households or multiple people.” Due to this lack of clarity, metrics such as reach and frequency – simple, straightforward measurements on linear – become obfuscated on CTV. As a result, marketers can’t be confident that their ad dollars aren’t going to waste.
Beyond the complete lack of common frameworks, ad measurement in CTV is challenging for more obvious reasons. Even with effective targeting, there’s no guarantee that consumers will watch an ad. They may get up to use the restroom, scroll another device or chat with their family members. Plus, if users are served ads from the same campaign across different channels, it can be nearly impossible to confidently measure multi-touch attribution – identifying what specific ads and what specific actions contribute to a given consumer behavior like converting.
Jason Fairchild, co-founder and chief exec at CTV platform tvScientific, agrees that this issue of divided consumer attention impacts marketers’ ability to measure attribution with accuracy. “We live in a world where the last click from search or social claims all of the attribution credit, but we’re seeing a different truth emerge as CTV advertising is scaling up,” he says. “Users see a CTV ad and respond via a second screen. This ‘second screen behavior’ often leads to a search, but it was not caused or inspired by the search – the search was inspired by the CTV ad.”
In search of a solution
To industry players, it’s clear that things need to change in more than one way. Many believe the solution is making CTV look more like digital advertising. “With the majority of TV viewing happening on streaming services versus linear, we don’t need to retrofit the new era of TV with a panel-based measurement model from the 1950s,” says Fairchild. “I think the new era of TV measurement will be very similar to search and social, where we can leverage sophisticated digital targeting on a one-to-one basis and measure actual outcomes instead of outcome proxies like delivery of reach and frequency against target demos.”
The sentiment is one echoed across the industry. And it’s the reason for a recently announced strategic alliance between Roku and Nielsen. Announced in March, the deal brings automatic content recognition (ACR), a real-time ad detection technology, and dynamic ad insertion (DAI), the ability to serve relevant ads to targeted audiences in real time, to streaming on Roku. The device maker says that, using Nielsen’s frameworks, ad performance and measurement will enable smarter, live buying and give marketers opportunities to monetize audiences in new ways beyond basic demographic information.
Louqman Parampath, the vice-president of advertising product management at Roku, explains that this new partnership builds upon a long history of the two working together to improve ad measurement. “Very early on, we integrated with Nielsen for what is known as dollar measurement, which equates to digital ad ratings. This is to say that for every campaign that you buy from Roku, we can tell you the age and gender bucket in which audiences who got exposed to that campaign fell.”
Roku continued to work with Nielsen to replicate Nielsen’s gross rating points measurement on CTV. Amazon and Hulu quickly followed suit, adopting a similar approach. In order to maintain a competitive edge, Roku felt it had to go further. “There are other kinds of questions that need to play out here, such as incremental reach,” says Parampath. ”If you buy across multiple platforms, how can you actually duplicate between users who saw on Hulu and also on Roku and also on Amazon Fire and so on? That is where a more broad currency is possible.”
This currency could come in the form of a unified identifier, which would enable marketers to more seamlessly share data and work across different platforms within the ecosystem.
Innovid’s Chalozin, like Parampath, sees a growing demand for a universal identifier in CTV. However, unlike some who says the ‘big four’ will continue to duke it out, Chalozin believes that it could be a real possibility. “[The industry] will create a unified identifier that will allow us [greater access to user data],” he says. He claims that the establishment of a universal identifier will reduce the fragmentation of the ecosystem and enable marketers to better calculate the value of impressions. “It’s in the best interests of almost everyone in the industry.”
Chalozin readily admits that any viable universal ID solution for CTV will need to be extremely privacy-centric. Today, 84% of Americans feel they have little to no control over the data that companies or the government collects about them, according to a Pew Research study. Combine this distrust with the aggressive shifts toward increased data privacy in the tech and legislative arenas, and it’s obvious that any CTV solution will need to stay the same course.
Little Dot Studios’ Swallow agrees that privacy needs to remain top-of-mind. However, unlike many other industry players, he says that giving marketers access to more granular and more accurate data is not necessarily an inherent good or a problem that even needs solving. “Yes, an entire industry has developed off the back of what has been termed ‘surveillance capitalism’, but that doesn’t mean we must be on an inexorable path to more granularity. Globally regulators are putting the brakes on this through the likes of GDPR and CCPA.”
Even so, Swallow says there is need for standardization. “For basic and anonymous measurement, there needs to be an independent standard with an agreed methodology. It happened at the birth of advertising on the web, where first-party ad serving gradually disappeared and was replaced by audited measurement and ad-serving tools. CTV and streaming need the same to give confidence to the market on the reliability and veracity of numbers.”
For many marketers, accuracy and granularity will remain top priorities. Chief among their concerns remains multi-touch attribution – and, more generally, simply understanding where consumers’ eyes are and when. Solving these challenges will require that key players develop more effective cross-media measurement capabilities.
According to Kimberly Gilberti, the senior vice-president of product management at Nielsen, the marketing research firm is working to ensure that “computation is being done in a way that is truly comparable” across channels. “For example, television today is measured on the average minute basis, whereas digital is measured sub-minute,” she says. “So even though you can put TV data and digital data together, their underlying data collection and their underlying calculations are actually very different. They are being put together in ways that are not truly comparable. And that’s really our mission with Nielsen One: to put these numbers together in a way that they are as ‘apples to apples’ as we can make them.”
Parampath says that this is one of the key focus areas of the Roku-Nielsen deal. “We’re working to measure reach across multiple streams – it could be linear, it could be connected TV, it could be a desktop and it could be mobile. As a user, you could get exposed to the same Pepsi campaign across all those channels on your mobile app, on your computer, when you watch traditional TV through cable or a set-top box, or when you’re watching connected TV. But as it stands, you don’t actually get credit for reaching the same person across those different channels.
“You want to be able to actually say that you got incrementally new users on these channels that you did not get on some other channels.” He says this is the central objective of cross-media measurement.
Ultimately, CTV needs to see unification of some sort. Without some agreed-upon frameworks or standardized currency, the ecosystem will continue to splinter, making ad measurement ever more uncertain. Across the industry, initiatives that seek to standardize data-sharing with qualified measurement partners are under way. And outside of one-to-one partnerships such as the Roku-Nielsen deal, major industry groups such as the Association of National Advertisers (ANA) and the Video Advertising Bureau (VAB) are working to enable cross-platform measurement, which could give marketers greater, more accurate visibility into metrics like reach and frequency across channels.
But how we shift the paradigm is a bigger question. Advertisers are sure to benefit from greater cooperation and shared methodologies – it’s critical that the right economic incentives are in place to help stimulate this unification.