Programmatic CTV Should Be Content-First, Not Audience-First

By Andrew Rosenman

"Audience-first" was in some ways a reaction to "content is king." A decade ago, as the programmatic space was reaching critical mass, the industry glommed onto the glory of cherry-picking audiences based on fancy new algorithms that allowed us to "personalize" impressions.

In so doing, many threw the proverbial baby out with the bath water by dismissing the importance of content and the brand value of cable and broadcast networks. That pendulum swing worked okay in a desktop/mobile digital marketplace, but as we move towards a programmatic era in premium CTV, less so.

As someone old enough to remember the 1980s, I think about the rise and eventual disappearance of "Captain Outrageous" Ted Turner, whose eponymous Turner Networks began the cable TV revolution. Turner pedaled the "content is king" mantra louder than just about anyone, and ultimately, he may have been more right than wrong.

TV audience reach planning starts with content first. Agency TV planning workflows, refined over generations, deliver a basic channel and series selection on behalf of advertisers. Once completed, audience segmentation algos can then add value and refinement to video investment decisions. Those of us in the CTV space would be wise to adopt this hybrid approach.

In other words, cherry-pick the right audiences watching programming that matches the tone and temperament of your client brands.

Instead, all too often digital planners start with a single principal input, such as men between the ages of 25 and 40 with $150,000 HHI. Their supply partners then magically conjure a bunch of inventory for open exchange or private marketplace (PMP) bidding, with scant regard for the originating content creator. To be generous, cookie-based digital audience projection has always been a bit of a dark art. These media plans can be twisted to help meet volume and revenue commitments made to large publishers.

A true and recent example was a financial services marketer who asked for a video targeting plan around its portfolio of individual retirement products. What came back was a recommendation to place its cute little brand mascot adjacent to horror, thriller, and suspense programming during the broadcast overnights.

Now it is entirely possible that people are up late at night worried about their retirement finances, and it is simultaneously possible that a segment of these worried consumers are fans of the horror genre.

But is it really likely that this audience is more likely than others to be predisposed to these particular retirement products or that while watching horror movies, these consumers are also chewing over the nuances of financial planning?

This is what can happen when "Big Algo" has too much influence over disciplined media planning. A magic black box has suggested who the segment is and where they can be found, and without applying any type of qualitative audience insights of its own, it simply presents this as an assumptive truth.

Nobody makes long considered purchases or switches banking and retirement providers in the middle of the night and asking the brand to be adjacent to chainsaw massacres and disembowelment isn't exactly a brand-safe recommendation. A robotic planning approach to what is essentially an emotional medium is not likely to deliver desired outcomes for clients.

A More Thoughtful Approach


Start with understanding the brand goals, which means taking the time to do the requisite homework. The inability to look at what a brand is doing from a 360-degree perspective outside of digital channels is a huge blind spot in programmatic. It's slightly delusional to believe that a digital media plan is independent from whatever other marketing activity is going on.

Focus first on the programming because that's where consumers start. Within the base of viewers, we all have signals that give us directional (and sometimes deterministic) information that allows us to assess the viability of the programming to deliver a strategic target at scale. From there, we can apply the segmentation to ensure that we are efficiently buying that programming, versus the programming "falling out" of the segmentation run. It's like ordering a double-filtered bourbon, smoother and higher quality than the rail liquor being poured at the local saloon.

Also, let's acknowledge that individual platforms will deliver variances within these audiences. An AVOD watcher on an OEM TV platform is different from an Apple TV device owner (who may be more selective and/or harder to reach), making Apple TV inventory worth more to an advertiser than the broader Roku audience.

Thus, it is critical to deliver inventory from the publisher that is not duplicative of the platform's same inventory sources. You will typically get deeper and more consistent metadata signals from the publishers; in order to manage frequency, you must see across multiple platforms upon which a publisher's inventory resides.

Of course, sourcing inventory from aggregator platforms is important as well, since there is so much of it, but the reconciliation of those impressions with the directly purchased publisher inventory is a necessary process that can prevent the cardinal sin of in-pod spot repetition.

Taking this "content-first, audience-next" approach will gain converts as brands continue to demand greater accountability of CTV campaigns. As more and more buying shifts from broadcast, let's bring along those practices that have helped to build brands for decades and bind them with our new programmatic methods to do even better for our customers.


The views and opinions expressed are solely those of the contributor and do not necessarily reflect the official position of the ANA or imply endorsement from the ANA.


Andrew Rosenman is Equativ's product marketing lead of CTV/video.