Way back in 2013, I got into a fight on Tumblr with another user complaining that Marvel movies included too much product placement. They cited the billboards shown at the end of the first “Captain America” movie, which I argued didn’t count because the scene takes place in the middle of Times Square.
In hindsight, that user might have been trolling. (They didn’t mention the fleet of Acura sedans that drives into the scene, which is a much more credible example of shoehorned product placement.) Back then, most TV and film productions didn’t bother with replacing the existing out-of-home ads in major metropolitan filming locations, except maybe to add an easter egg or a bit of in-universe set dressing.
Technology’s come a long way in the last decade, though. And for some marketers, the billboards that appear in the backgrounds of shots – or, heck, even just the blank walls of a building – feel like prime real estate for even more ad placements.
That’s the product that AI startup Rembrand is trying to build, more or less. The company’s goal is to make programmatic, dynamic billboard insertion a viable model for TV content owners and advertisers, on the grounds that it’s less intrusive to the viewer experience than the typical commercial break.
We’ll fix it in post
Rembrand’s focus is on creating ad inventor that appears directly inside scripted content. It focuses on placements that won’t interrupt the story but will still appear prominently enough that brands will be able to measure impact, the company’s chief product officer, David Wiener, told me.
Establishing shots, so called because they set up the location, time or context for a scene, tend to be a pretty great fit for Rembrand’s digitally inserted billboards, Wiener said. Sometimes, that even means inserting new shots into existing programming to create more available inventory.
For example, say the Hallmark Channel is airing a TV movie set in Chicago, and there aren’t any shots where it would make sense to add a billboard. Rembrand can instead use AI to generate an establishing shot of a building exterior, complete with a canvas to contain ads.
The process is a lot easier said than done. Anything added into the scene has to match the color and tone of the original and can’t be too fake-looking or incongruous to the time period – no cell phone ads in a TV show set in the ’60s, for example.
All of these factors contribute to why the process isn’t programmatically transactable at scale yet. Although Rembrand has started doing pilots to test out its programmatic capabilities.
So far, Rembrand has encouraged the publishers it works with to treat the ad inventory like a direct sponsorship opportunity, where both sides of the transaction know exactly when and where ads will run, Wiener said. “Nothing ever goes out that a brand obviously doesn’t want to be in, or that a content owner is surprised by.”
However, it’s also worth pointing out that while content owners are involved in these deals, the actual creators of the content – directors, set designers, cast, crew, etc. – likely don’t have any say in what gets added into the production they worked on. Which, to me, feels a little unsettling. But, hey, that’s show business, I guess.
Knock it off, I’m watching something!
Here’s the most important thing, though: Does it work? Do audiences actually respond to these kinds of stealth ads?
According to Rembrand, its in-content advertising integration performs better against a number of metrics when combined with traditional TV spots, including purchase impact, net attention reach, brand awareness and overall favorability.
Rembrand’s results do seem to adhere to the common consensus among CTV advertisers, which is that emerging in-content formats (like squeezebacks, overlays or even home screen ads) perform better overall than commercial breaks.
Some, like Wiener, attribute this effectiveness to the idea that new formats are inherently less disruptive than regular ads, given that the content doesn’t stop running while the ad does.
Personally, I’d push back against that framing. Anything can feel disruptive to a sufficiently distractible viewer – heck, some people can’t even watch TV with the subtitles on.
More importantly, in-content ads aren’t skippable, and they don’t announce themselves as obviously as ad breaks do. Meaning that viewers are less likely to be doing something else, like getting up to use the bathroom or staring at their phones.
But a lot of people, like that Tumblr user I argued with, can have strong reactions to the appearance of any brand messaging inserted into their entertainment, regardless of context.
As these dynamic ad formats get more ubiquitous – and as they get more programmatically available, too – brands and advertisers will have to decide where best to draw the line. How many ads is too many? When does an ad drawing a viewer’s eye away become a bad thing?
In the meantime, keep an eye on those billboards in the background of your favorite TV show. Someday they might be unique to you.
Questions? Comments? Wanna reminisce about the old days of Avengers Tumblr fandom? Drop me a line at victoria@adexchanger.com.
