OTT Ad Tech's Biggest Challenges (and a Few Solutions)
The digital consumer's experience in the near future will enable viewers to claw back personally identifiable information, as third-party cookies are eliminated and first-party tracking must be opted into. This will leave the ad industry in a tizzy figuring out how to continue to do targeted ads. (Is content personalization next?) Previously, trying to navigate online without leaving little pieces of your personal information all around the web was very, very hard. It was like negotiating with a brick wall: "If you don't give us access to your personal information, we can't provide what you're looking for." Where to put your focus now should be top of mind.
Future Targeting 101
"Identity is an under-discussed topic in our industry," said Magnite CTO Tom Kershaw at IAB.ALM 2021 in March. He has several approaches for ad targeting. "The first is login, which can be used to create a persistent cross-party identifier." This is what I like to call "give me your email." Two groups are at work on this. The Trade Desk has created Unified ID 2.0, an initiative that relies on email addresses from consumers who give their consent. And LiveRamp has developed Authenticated Traffic Solution (ATS), which uses consumer opt in for data to be used by publishers and advertisers. Both groups are working to create some kind of identity verification standard, and when one becomes more dominant, the other may or may not say uncle.
Kershaw's second approach is contextual-based advertising, which shows ads that are related to subject matter. If you watch a program about travel, you see hotel and credit card ads. It's not what digital advertisers have been using, but given the alternative of having absolutely no data about a viewer, this has its merits.
The last approach on Kershaw's list is my all-time favorite. "In our industry, the most important first-party data is really playlists," he said. "If you look at consumption levels on a [connected TV] platform of different types of video content, those signals are super valuable, and you don't even need to go too much further beyond that. People think of first-party data as being very granular, and it's not. It's what TV shows have you watched in the last 30 days." This seems like a very good, trustworthy, easy-to-understand, and non-evasive approach. Plus, I'm guessing consumers would be happier to know that all you want is to track the shows they watch and possibly their email address, as long as you promise this information is only going to provide them with content, ads, and nothing else.
The World According to Google
Google floated a proposal in March 2021 that set the ad industry on edge. It plans on tracking cohorts using "a semantic classifier to the Chrome browser. It will scrape page content and URL data and qualify users into small segment groups based on their navigation history. Everyone's still being tracked, but now only by Google," writes Pierre Diennet, VP of product management at data solutions company Lotame, in his article, "4 Proposals for Navigating Post-Cookie Identity."
As of now, there's no information on if this proposal will be shared outside of Google. Also, in April 2021, The Wall Street Journal reported on a Texas antitrust lawsuit that claims Google was using its knowledge of sales prices on both the buy and sell side to goose its business. The same lawsuit uncovered more information about Jedi Blue, a confirmed data deal to ensure that Facebook was able to identify 80% of mobile users and 60% of desktop (excluding Safari) users. This sounds suspiciously like a bully is kicking sand in the other kids' eyes and running away with their data.
Ad people say things like this statement from a Google article: "Audience measurement will need to integrate traditional, panel-based gross rating points and internet-enabled, census-based measurement before the potential of addressable advertising can be fully realised." And such language is common to all ad people, not only those from Google.
Translation: Today's measurement needs to combine the existing measurement based on audience estimates collected by Nielsen (from traditional broadcast) with analytics collected from digital viewing before ads can be targeted on broadcast content. Previously, this was a big problem, but now, the laser-beam focus has moved on.
Also, here's a refresher for some terms so you don't have to run to look them up: contextual (what the viewer is looking at), authenticated identity (deterministic), and non-authenticated identity (probabilistic).
Rethinking Data Strategies
Previously, first-party data included things like age, income, gender, interests, device-viewing details, IP address, etc. Now, the idea of creating a new and improved first-party data strategy should be top of mind. More data doesn't equal better intelligence, but the right data strategy does. Who owns this?
"A company's digital marketing arm would be the best group to set the strategy," says Louqman Parampath, VP of product management for advertising at Roku. This is most appropriate for ad targeting; however, there will also be other targeting going on with recommendation. (I'll go further into this in a future article about centralizing viewer analytics.)
If you're lacking sufficient first-party data, one resource to go to for authenticated data is the vMVPDs, if they will share. "Platforms with first-party relationships provide resources that enable consumers to better understand privacy policies and product features to give consumers more control over how their data is collected and used. By contrast, companies that have no direct relationship are both more limited in the kind of data they can collect and the information and control they can offer to consumers," says Parampath. The key takeaway is that you likely need a data strategy refresh.
Creating Something Out of Nothing
I'd like to add my own idea: In the six-degrees-of-separation world, advertisers are missing the point. If I saw an impressive advertisement, even though I don't need a new car, cat litter, or motor oil, I would surely tell someone who did. In fact, I likely would tell as many people as possible, which is why YouTube, TikTok, Twitter, and everything else in social media have thrived. So, the formula is to create great ads, and word of mouth will amplify it, even with poor to no targeting.
Advertising has bigger problems: It gets in the way of itself by using way too much jargon, consumers don't trust it, and there seems to be an inability to play nice.
People like free content, so figuring out better ways to serve advertising should be the key problem every media company is solving. The Roku pause ad is wonderful. It's done well for the company, and it has created inventory where there was none before. Each time you pause the video, an overlay can pop up. TheTake has combined watching and shopping into one application in which ecommerce meets streaming. Ryff is slipping 3D images into previously blank spots within a piece of content, with the content creator's permission. These are three very innovative ideas. Our industry needs to come up with a whole lot more.
While 15s and 30s are the bread and butter of the ad industry, I challenge technologists to spend more time thinking up ideas that will scale under the new privacy requirements. Less jargon and more innovation should create some truly brilliant, business-altering ideas so we can get back to watching something good and stop seeing ad executives shoot themselves in the foot.
[Editor's note: "Ad Sense" is a new column that will appear in every issue of Streaming Media Magazine.]
As ad-supported viewing grows, the industry grapples with three key challenges: privacy, workflow fragmentation, and scale
Here's how five industry experts are finding their way through the complicated world of advertising technology.
It's time to move beyond our outdated, Web 1.0 approach to advertising and embrace stateful ad podding. The future of connected television depends on it.
As ad-supported OTT services grow, the video engineering world needs to understand the complicated advertising technology landscape.