Magnite Team
7 min read

IRIS.TV Explains the Rising Importance of Contextual Targeting for Digital Video

Note: This post originally appeared on the SpotX website.

This year has accelerated the development of several digital video trends. Among them, we’ve seen audience data strategies evolve in response to new privacy regulations and the death of third-party cookies. We’ve also seen a surge in video viewership across all devices, including connected TV (CTV). 

Sean Holzman, Head of Ad Platform Strategy at IRIS.TV

As a result, it’s becoming increasingly important for media owners to help buyers implement category-based targeting strategies across their inventory. IRIS.TV, a video intelligence company building the underlying API infrastructure for video data, is making this possible by connecting disparate data sets from a media owner’s content management system (CMS) and video players through its chosen supply-side platform (SSP) and demand-side platform (DSP).

I spoke with Sean Holzman, Head of Ad Platform Strategy at IRIS.TV, about how media owners and advertisers can improve targeting and brand safety and suitability for video across screens.

Nick Frizzell: Tell us a bit about IRIS.TV and what you’ve been up to lately.

IRIS.TV Contextual Video Marketplace

Sean Holzman: IRIS.TV is a video intelligence company known for our Video Personalization & Programming Platform built to help publishers and broadcasters increase video consumption, engagement, and retention while enabling editorial programming controls.

With that said, a lot of our current focus is on our new Contextual Video Marketplace which acts as a neutral platform to enable contextual and brand-safety data targeting on cross-device video for the very first time. The underlying personalization technology puts us in a unique position in the market to service this need, and we are seeing rapid adoption across both the supply and demand sides of the market. This is helping our publishers increase fill rates, effective CPMs (eCPMs), and overall revenue, while enabling the buyers the transparency needed to see what content they are running against, and increase campaign performance and ROI.

NF: How have the changes looming with data regulations, cookies, and device IDs affected how advertisers target their campaigns? What do players across the ad ecosystem need to do to adapt to our privacy-first world?

SH: This is a huge topic right now and rightly so. The digital advertising ecosystem is going through a massive transformation around how campaigns target and reach audiences, which today is primarily audience-based.

We really have two things happening at once; changes in global privacy regulation and a huge increase in video consumption across all devices. 

With that, there need to be new and differentiated ways for publishers to represent their inventory and for buyers to target their campaigns. We are advising publishers, SSPs, DSPs, and media buyers to all adopt a contextual strategy that can complement their current and future audience targeting capabilities, whatever they may be. 

Contextual targeting has been around for a long time, but has fundamentally been based on analyzing the text on a web page. With our Contextual Video Marketplace, buyers and sellers can now target campaigns based on data pulled from analyzing each individual video, for both short-form and full-episode formats. 

NF: How does IRIS.TV achieve brand-safe targeting for individual videos rather than at the page level? How does this approach extend to CTV inventory?

SH: Contextual targeting for display relies on crawling the text on a page, which is publicly available. Video data is not publicly available and there is no “page” to crawl for CTV. While there may be business reasons why a publisher would want to limit access to video data, the main roadblock to accessing and analyzing video content at scale is technological. Video is locked behind a publisher’s CMS and video player and is inaccessible without a direct integration. In addition, the video metadata and content feeds are not standardized so it presents an enormous challenge for data providers to integrate, ingest, normalize, analyze, and segment assets. 

Prior to integrating their solutions into our marketplace, contextual data companies were only able to analyze page URLs. This prevented contextual and brand safety platforms from unlocking the most valuable video inventory for targeting on the open web and completely locked them out of CTV.

IRIS.TV has spent the last seven years building video data infrastructure that integrates with every video player and CMS, and can ingest and process any content feed. This provides all the contextual data providers that integrate with our Contextual Video Marketplace access to video so they can apply their unique segmentation. It is akin to getting access to the YouTube API, but for the open web and CTV.

Until now for premium online video, data providers could only use page-level information to make a best guess on the context of an embedded video. With the increased emphasis on brand safety, suitability, and transparency, this stop-gap solution is no longer acceptable. 

For CTV, there is no page-level data to analyze, so this is a first-of-its-kind offering that is truly bridging the gap between digital and linear TV buying.

NF: How can media buyers still reach their target audience by using contextual segments as opposed to data segments?

SH: It doesn’t have to be an either/or question. Contextual targeting is complementary to all forms of buy-side targeting. But if one is choosing between one or the other, contextual targeting does provide many benefits that can drive better results to behavioral targeting. In a study conducted with FOX, CTV video was shown to have multiples of return on ad spend (ROAS) compared to all other video and digital media. The study also showed that when contextual targeting was applied, the brands that participated realized a 35% sales lift over ads when no contextual targeting was applied. A study by GumGum and Dentsu revealed cost efficiencies, with campaigns achieving a lower eCPM, CPC, and viewable CPM (vCPM) when contextual targeting was applied.

NF: What does the shift to contextual mean for media owners?

SH: For media owners, it means that their inventory is more valuable and will command a higher CPM. There is a greater emphasis on alignment with the premium and high-quality content. Research is continuing to show that consumer perception of ads and brands is influenced by the quality of the environment. When it comes to CTV, the lack of transparency has limited the investment in ad spend. By contextually enabling CTV inventory, publishers can provide transparency without giving away sensitive content data for both legal and strategy reasons. 

NF: Fraud and inventory quality are always top concerns in video. How can media owners ensure a brand-safe environment while also helping advertisers improve their targeting and relevancy?

SH: Every video onboarded onto the contextual video marketplace is given a persistent global unique identifier called the IRIS.TV ContentID. This solution unifies thousands of integration points in the advertising landscape into a single data stream. The result is that marketers can now reach viewers at the exact moment they are watching video. This will increase the relevance, effectiveness, and overall impact of their brand messaging and advertising. This is all achieved by analyzing the content of the video itself (not just scraping the text on the page) while protecting the privacy of end users, and does not rely on cookies. 

NF: News content often is an avoided contextual category and deemed too risky to buy. How does IRIS.TV recommend approaching news content? Is there an approach that allows a media buyer to run on news publications, but carve out the exact segments or pages that are most suitable for their brand?

SH: News is a great place to reach consumers. The challenge is, how do you truly avoid brand-unsafe or unsuitable content without eliminating all content? The traditional way to solve this issue was domain and keyword block listing. This can be heavy-handed and controversial. Early solutions would claim to know the difference between “shoot a gun” versus “shoot a free throw.” But as many reports to the contrary have surfaced, brands have been reluctant to engage.

If a picture is worth a thousand words, then a video must be worth a million words. What we are able to do is provide the leading AI- and ML-based data solutions with access to what essentially becomes massive training sets of video data so that their algorithms can truly understand something that is not brand-safe or suitable. For example, a brand may not want its ad to be aligned with a video titled “Hundreds of Nursing Home Residents Die of COVID-19,” but would be more comfortable with being associated with “Ten Ways Parents are Managing Work and Home Schooling During COVID-19.” 

What we are seeing is that these systems are getting smarter. So much so that the percentage of videos flagged as brand unsafe is going down even though there is more news video than ever before. What’s more is that positive targeting of categories such as finance, sports, and lifestyle are being unlocked in ways that are really making a difference. Brands are now targeting content related to topics such as social justice and diversity, which previously had not been possible to achieve at scale.

NF: Media owners have faced numerous challenges this year. How can they increase fill rates and CPMs to close out the year with a successful Q4?

SH: The faster you can adopt a contextual strategy, the faster you will be able to see your CPMs and fill rates increase. COVID-19 definitely disrupted the advertising landscape, but buyers are getting familiar with contextual targeting for video and are leaning in aggressively into Q4.