By the end of this year, more than 68% of the US population will be CTV users according to eMarketer. But for advertisers, reaching the right people through CTV remains a massive challenge.
One of CTV’s strongest draws for brands is getting access to the same targeting methods digital advertisers use to reach ideal customers. But gaps between user and device persist for CTV in a way they don’t for digital.
CTV viewers often respond to an ad through their phone, which means the targeted IP address (the smart TV) and the one responding (the phone) won’t always match. Device graphs try to resolve this, but they’re flawed. Sometimes you just don’t know who your ad reached.
For example, CTV advertising targets households rather than individuals. This means an advertiser could easily target an IP address that’s been mapped to a 45-year-old IT professional and instead reach his 10-year-old daughter. Or his neighbor, friend, or child who lives at college. Many streaming viewers confess to sharing their account beyond their household—despite the crackdown by streaming services.
Plus, IP graphs change. No one’s confirmed exactly how regularly this happens, and so far there are no great solutions. Not to mention that multiple types of IP address exist. A given advertiser’s CTV logs may have a newer or older version of IP to which website tracking pixels can’t match.
Third-party data also commonly lacks transparency. Even inquisitive marketers may find themselves in the dark about how users are grouped into targeting segments or how frequently data is refreshed.
That’s just addressing the inaccuracies inherent to third-party, IP-based targeting. The cost is another matter. Even if you could be sure every individual you targeted was exactly who you wanted to reach, the cost might outweigh the benefits. Every layer of targeting equals more fees, meaning you need an even higher return on your campaign to make those fees worthwhile.
New privacy legislation and the decline of the cookie comprise the final strike against third-party IP targeting. The problem is ultimately a data challenge, and it’s only going to get worse. Already, according to Nielsen, just 23% of global marketers say they have the quality data needed to maximize their media investments. Brands should test alternative targeting methods, even if only to prepare for a possible future where the standard IP address goes away.
Contextual targeting is one such method that’s experienced a renaissance. By letting advertisers target based on content rather than the individuals watching, contextual sidesteps the fees that come with traditional third-party approaches while ensuring ad relevance. A sports drink is shown during a football game. A home goods store with a new line of kitchenware advertises during a popular cooking show.
But contextual targeting also faces challenges. In 2023, the IAB created a new video classification for accompanying content to making matching ads to related content easier, but there’s still a lack of standardization in how publishers label content for contextual buys. Some content isn’t labeled at all. This makes scale challenging, if not limited.
First-party data, of course, has been touted as the long-term targeting solution. The question is what to do with it. Obviously, retargeting has value, but limited reach. Any growth-driven brand should use CTV to attract new customers, not just reconvert existing ones. Look-alike audiences are another answer, but you still need third-party data to build new audiences based on the attributes of your current one.
The best solution might be an entirely new type of targeting. AI’s evolution offers the potential for targeting solutions that simply weren’t possible a few years back. One of these solutions is algorithm-based targeting that combines the best of contextual and look-alike audiences without paying to target specific user profiles or dealing with incorrect IP addresses.
Advertisers can use machine learning to integrate and learn from site and bid stream data including geography, viewing habits, app usage, daypart, and device type. Then, algorithms identify patterns indicating which factors correlate most strongly with ad engagement. These patterns then determine your media buys. With this approach, you avoid reliance on third-party data, expand contextual beyond genre-based buys, and learn as much as possible from your first-party data. It’s a new solution, but I’ve seen multiple advertisers already prove it outperforms their other targeting methods.
All this isn’t to say advertisers should stop using third-party targeting. But they should understand its limitations and actively explore other approaches to avoid over-reliance on an expensive, flawed, and still-evolving aspect of the CTV ecosystem.
Read more from VP Strategy Dan Cleveland and VP Analytics Matt Hultgren in this article covering the must-knows of CTV targeting strategy.