Resolving Challenges in CTV: Interview with Erhard Neumann, Spearad


In association with Unruly.

In this exclusive interview with ExchangeWire, Erhard Neumann, MD, Spearad, and head of advanced TV, Unruly, outlines the current state of play in the CTV ecosystem, and how current challenges can be addressed.


What are the key issues affecting the CTV viewer experience today? How have these evolved over the course of the past two years?

I would start with frequency capping, which is such a hot topic, particularly in video-on-demand, because we all know the experience of seeing the same ad over and over again. To me, it's more a problem of volume in that there are not enough quality ads available and therefore, no way to prevent sending out the same ad again and again. To avoid this, we may need to introduce other mechanics. One example could be, especially on AVOD, reducing the length of the ad break if the platform doesn’t have high-quality ads to fill it.

The other thing that I see is that many publishers rely 100% on third-party monetisation platforms without leveraging the benefits of an ad server. In many cases, they are trying to maximise their monetisation partners, but what can happen is the publisher sends one ad call and may get the same ad from four or five different demand sources, with no technology controlling the ad. The system sometimes doesn't even know that it sends the same ad repeatedly.

We believe we need to resolve this from the technology side, potentially via a unique ad ID, which means that whenever the ID comes in, we see that it has already been used, and then we can cut it off to prevent ad duplication. Using an ad server or an ad decisioning system that controls the full ad break, to know which ad is placed in the first position, second, third, and so on, is another way to solve this problem.


How can the industry work on a collaborative basis to ensure effective frequency capping?

User IDs and household IDs present a problem here. We don't really have user IDs for a CTV device in the living room, because many people are potentially watching the device separately or together. This is a difficult challenge to solve quickly, especially in Europe with GDPR. We also know some broadcasters renew the user ID every 24 hours, which limits the ability to frequency cap for longer than a day.

As previously mentioned, the first thing we need to do is, find a way to identify what ad is being served irrespective of which demand partner it comes from. Then we can decide on whether we should have this ad inside an ad break or not, as it may have already been used. Solutions are already being tested on the market, to identify the ad by mechanics such as reading frame-by-frame to identify duplication of the frames. However, this is a challenging thing to do in real time, as you need to take the ad in, analyse it, put it to one side while delivering something else. Once the ad is analysed you can insert it into the ad break but you can’t yet do it in real time as you can’t analyse it in under 30 seconds.

Secondly, publishers and broadcasters should implement a leading system for holistic ad decisioning and ad break management, one which can receive all responses and places, according to the numbers of rules set. One of these rules of course being frequency.


What are the key considerations for publishers when ensuring their CTV infrastructure is maximising viewer experience? What efficiencies can modern CTV ad servers bring here?

Erhard Neumann, MD, Spearad, and head of advanced TV, Unruly

Modern ad servers are built for the transition from traditional linear TV to addressable TV, enabling greater control over the ad break behaviour so that it’s filled in an optimal manner and facilitates a high-quality user experience.

For CTV, it is very important that the viewer receives a traditional TV-like experience – with quality content, ease of use and relevant, non-glitchy advertising, particularly as the ads are not often skippable. To this end, publishers often don't use an ad server to its full extent, they just use it to connect to monetisation partners. That’s what they need to rethink, to use an ad server or ad decisioning platform which helps them to set better rules so that the ad break is filled far more effectively and efficiently

Advertiser competition is also something that influences the ad break behaviour. On a website you can go on content categories. In the CTV space, you definitely need to have advertiser rules set, such as distance rules instead of category-avoidance rules, in order to make sure you get the same kind of TV-like experience known from traditional TV.

One more thing that publishers should also be aware of is, while there is already a huge audience using CTV, if you add more than two types of targeting criteria, the size of the audience rapidly decreases. I have often observed an overemphasis on targeting granularity, and brands should be aware of the effect of their campaign reach.


How should brands direct their creatives to engage CTV audiences? How can they work with publishers and technology partners to ensure their content is customisable and optimised according to specific audiences?

Pricing is not the only rule deciding this, it could also be that some types of ads, especially targeted ads, find a better position at the beginning or end of a break. It's a combination of determining the business objective of the ad break, and a mix of rules for things like price, competitive separation, ad slot position and targeting the content owner.

Traditional TV relies on panels to identify audience reach throughout the day, while in the CTV and AVOD, similar to digital, ads are served to relevant viewers as they engage with content.

When a user is watching a movie via AVOD for example, there's a one-to-one connection in which the user is much more sensitive to how many ads are streamed within the AVOD content, the length of the ad, and the relevance. With this one-to-one connection, publishers can be more customised with the ad content and flexible when it comes to the length of ad breaks, to avoid annoying the viewer. If you don't have enough ads available, just make the ad break shorter – Unruly's ad server can do that!

To optimise the ad experience, it may also be better to show two ads from the same advertiser but in a way that doesn't feel like it's duplicated. You start a story with one ad, and you end it with a second, and in doing so, you tell a story.

Finally, we have a big problem again with the identification of users. You can use meta information that often comes with the content, such as genre, actor, products and language, in combination with data available from households. The same is when you can identify contextual parts of the content and translate them into keys and values, making ad selection for an ad server much more precise than just "targeting a user." It's a mix between targeting the user, using meta and contextual information from the content, to deliver the right ads.


What effect will identifier deprecation across mobile and display have on CTV viewer experience, and how can any potential negative effects be mitigated?

I think the industry should see the deprecation of traditional identifiers as an opportunity to do things in a more consumer-centric, privacy-forward manner.

When it comes to CTV, sharing metadata and contextual information is key, as are traditional TV metrics such as about time of the day and where and how to best reach the audiences. This information can be all put together to create a relevant ad experience for consumers.

It’s perfectly acceptable to give up some precision on the user level, as long as the quality of the advertising and the user experience is advanced, which is what modern ad servers can facilitate.