He contrasts CTV with linear, arguing that an aspect which makes linear attractive for advertisers is the transparency around what content ads are being placed next to. Hyden remarks that linear has traditionally been bought and sold based on audience data (provided predominantly by Nielsen) and data around the content provided by broadcaster networks.

In a linear environment, Hyden says: “You know who you’re reaching and what kind of programming you’re running adjacent to. You can ensure it’s safe and suitable for your brand, measure the impact of a campaign over time and reallocate budgets in locations you know performed best.”

According to Hyden, the reason CTV is not transparent currently is not because media sellers are hiding anything but because the infrastructure needed to make it transparent had not existed previously; with CTV experiencing explosive adoption over the last 18 months, the industry is “playing catch up.”

IRIS.TV recently partnered with Publica to integrate video-level contextual data into the latter’s CTV ad server. The company says the solution ensures brand safety and transparency while “dramatically streamlining the process of introducing contextual data for planning and targeting advertising campaigns.”

Hyden says: “There are two main ecosystems at play: the advertising world and the video streaming world. Those two worlds haven’t really spoken to each other in a long time. What we do is connect the video infrastructure with the ad tech ecosystem that buys ads against that video content.”

He clarifies that contextual advertising has been a feature of digital for a long time. With display for example, the ecosystem operates by using website URLs that are passed into the bid request between buyers and sellers. These website tend to be open domain, and as a result companies over the last 10-15 years have been able to build technology which scrape webpages’ texts to decide if they are brand safe, and determine the topical nature of the content is. Through these pre-bid integrations, DSPs are able to parse the text on a website and tell buyers whether the content is safe to advertise alongside.

Hyden remarks: “Video doesn’t work that way. Video URLs are not parsed in the bid stream. Video URLs are not available in the ecosystem for anyone to analyse and scrape because that kind of content is very expensive to produce.”

He explains that the role IRIS.TV performs as a video data platform is to “sit in the middle” and ingest all the video data coming from the video URL, as well as the metadata coming from publishers.

“That’s what we do,” Hyden says, “We work with publishers to ingest their whole video library. We clean up that data and then provide that data to data companies like Comscore, GumGum and many others, so they can analyse those videos and create scores for that content in a very similar way that would be done for online articles.”

According to Hyden, the integration with Publica represents the “last missing piece” in enabling contextual advertising in the CTV space. Publica is the first ad server to activate IRIS-enabled data for direct deal forecasting and targeting. Hyden says this takes the operational strain off of publishers: “They can just log into Publica to set up a direct deal like they would any other deal. They click on their contextual segments, select from a neat drop-down list of values and then launch the deal.”

In addition to ensuring brand safety and making CTV transparent, Hyden believes the solution also improves the efficacy of campaigns. Without knowing what content their ads are running adjacent to, it becomes difficult for advertisers to determine the content they should advertise alongside for optimal performance. He says: “By understanding the type of content running adjacent to the ad and being able to control for those differences from a targeting perspective, you can align your spend against verticals that you know will perform well.”

She said: “Unless you are a very specific advertiser, you can’t build scale [through targeted CTV campaigns]. We have got to collaborate more. If you are a big brand advertiser, you’ll be using this as reach extension…and that comes back to needing centralised measurement, and being able to buy from one point in the same way across all different sources.”

She also believes that the costs associated with buying addressable spots on quality CTV inventory is a large barrier for some mass market brands. Coteman elaborated: “Why some mass market brands haven’t been leaping on the CTV bandwagon is about CPMs and CPTs – on [linear] TV, historically, they’ve been paying quite reasonable CPTs. Quality CTV inventory is much higher priced.”

Coteman was joined on the panel by Chris Edwards, Director of Business Development – EMEA at Rakuten Advertising, Paul Gubbins, VP of CTV Strategy, Publica, and Jaidev Kakar, Director of Advertising Solutions, CTV & OTT at PubMatic.

Gubbins argued that when search and discoverability functions for the AVOD/FAST space improve, viewers will spend more time within that ecosystem: “That’s when we get inventory liquidity and that’s when we start to get to a particular point around scale when it becomes really attractive for those big linear advertisers to transition from those traditional environments into addressable environments.”

Kakar said the low cost of programmatic buying has opened up space for certain direct-to-consumer and local brands to advertise on connected TV. He said: “We’re seeing studies which show – probably because of the type of content or audience or because they have more local reach – that when [these DTC and local brands] buy with broadcasters there’s more mass, but there always seems to be around six times greater ROI through programmatic than linear.”

He suggested that smaller brands may be better positioned to see their ROI from advertising on CTV while larger brands will find it more difficult, needing to conduct more econometric modelling to see the effect of advertising on the platform.

Agreeing with Coteman about the challenge posed to CTV by measurement, Gubbins noted that brands are struggling to measure their reach holistically across a growing number of AVOD and FAST services. In particular, he believes ‘legacy’ linear advertisers may only now be starting to test advertising on BVOD and only because they can still apply traditional TV measurement practices in that environment.

While Gubbins does not believe there is adequate cross-platform measurement yet, he thinks initiatives such as C-flight and Project Origin can help advertisers try to understand incremental reach, attribution and frequency management when they plan and buy in TV environments.

Gubbins goes on to predict that measurement for CTV will differ from measuring traditional linear. He remarked: “Traditional TV has taken a panel-based approach to measurement. With the evolution of addressable TV and connected TVs that are plugged into the internet, it’s now more about impression-level measurement. The way we’ve historically measured for traditional TV may not be the way we do it moving forward because fundamentally – while it’s still a big screen TV in people’s living rooms – the delivery mechanism is very different to what it was maybe 15-20 years ago.”

Edwards registered his belief that there will be a universal approach to CTV measurement which the trade consolidates around, but there may be a long way to go before this is achieved.

Another issue which the panelists believed was important to support further monetisation of free connected TV, is achieving a “linear-like” quality to ad breaks in the AVOD/FAST space.

Coteman described broadcast linear ad breaks as “best practice” and highlighted some of the challenges CTV faces in trying to emulate linear. She said: “What people don’t understand is that it’s really difficult to run a streaming product. You’re trying to integrate lots of different proprietary and third-party pieces of technology that don’t all speak to each other properly, and not everyone has thought of every part of the process – especially when you’re approaching it from the broadcaster perspective. We’re at a stage where we’re all learning and developing and therefore, unfortunately, some viewers aren’t receiving the best service.”

Gubbins elaborated on these technical challenges, noting that AVOD/FAST services are relying on programmatic trading to monetise their ad-break, but that each of the different SSPs that participate in a programmatic auction for a slot within a publisher’s ad  break may be responding to that publisher in a slightly different way. This will often mean that bidders may not include the IAB category of the brand, or the name of the advertiser, which results in viewers sometimes seeing the same ad back-to-back in the same ad pods.

He said: “If advertisers can’t manage basic things – such as frequency and competitive ad separation – like they’ve been doing for the last 50 years on linear, they will spend their money with another streaming service that can provide structured ad pods.”