The demand-side platform (DSP) category was starting to get pretty boring.
Forrester stopped conducting DSP waves last year, due to the minimal differentiation and innovation. And VC investments have all but dried up.
But the programmatic ecosystem has geared back up as tech companies eye “the next battleground” of cross-channel video, said Philip Smolin, Amobee’s chief strategy officer.
The next generation of ad tech is already here, Smolin said. And in many ways, it resembles the advertising market that predated the internet, run by large incumbents and without user-level visibility.
AdExchanger caught up with Smolin about how Amobee and other DSPs have evolved to compete in this new market.
AdExchanger: How do DSPs have to change if they want win in television and connected TV (CTV)?
PHILIP SMOLIN: I think the very definition of what programmatic and DSP mean will have to change.
Ad tech providers – by which I mean programmatic acronym companies – were built on core assumptions that stemmed out of display advertising 10 or more years ago. One assumption was that advertisers would have consistent user or device-level tracking. There was also the idea of a near infinite supply of media that’s best sold through real-time auctions.
But many of these assumptions that underpinned programmatic no longer hold true in the television-converged media landscape. And consequently, a lot of the tech built by DSPs doesn’t work well for those campaigns.
How so?
For one thing, within connected TV there needs to be a recognition that a device represents a household as opposed to a user profile.
So take Amobee. We have a device graph that could map to a household graph, and we can augment that with smart TV panel data from an Inscape or Nielsen Gracenote. But what we get back from those TV data suppliers is the smart TV device ID and program exposures. We can see that they watched on a Roku device, but what we don’t get is Roku’s unique user-level ID.
I think what you’ll see – and what we are already seeing – is experimentation with more of a gated-garden approach.
Are there any promising examples of that approach?
Roku’s dataxu acquisition is a testament to this change. They’re going down the path of building their own walled garden, not unlike YouTube with Google’s DV 360 (Google’s DSP, Display & Video 360) as a front end.
So Roku in theory could allow media planners to target by individual user profile within a household, instead of going by household based on the IP address.
I think in the next year we’ll see varying levels of opening and closing access. So there will be walls, but maybe for select advertisers Roku will open the gates and allow more device ID data.
Ad tech companies should also do more to combine with panel-based strategies, like Nielsen, Vizio and others
I would expect programmatic companies to be pushing advertisers away from panel-based data.
Well, CTV is growing fast, but it was only 10% of overall TV spend last year, I think. That’ll grow to more like a third of the TV market over the next several years.
That shows how much ad tech players need to accommodate media strategies that existed before we did.
In some homes, there could be multiple connected devices, none of which are cookie-based or allow digital targeting – with recognizable device IDs available to advertisers. There are equipment manufacturers such as Xbox, PlayStation, LG and Vizio, as well as OTT devices like Roku, Fire TV, Chromecast and Apple TV, none of which speak to each other or can be measured holistically.
Because of this, connected TV’s potential for targeting and attribution of a quality that digital advertisers are accustomed to has been largely unfulfilled. It’s prevented many advertisers from taking advantage of the medium because doing real user-level, cross-screen campaigns and measurement is such a challenge, or not even feasible yet.
Recognize that panel data has its limitations, but it gives advertisers a way to understand whether or not they’re actually reaching their intended audience.
How does the advantage or disadvantage change for independent DSPs compared to owned-and-operated players like Google, Amazon and now dataxu with Roku?
It does change, but I think it’s also part of a broader maturation of ad tech that’s going on. But it is accelerated by the convergence with television.
The constant introduction of new media channels like streaming audio, out of home and, most importantly, TV and streaming video requires dozens or even hundreds of engineers to continue evolving the software platforms.
Economies of scale come into play very strongly in ad tech as well. Being able to process tens of millions of ad impressions per second requires millions of dollars in hardware and data centers distributed around the globe.
Ultimately, this means operating a DSP has an extremely high fixed-cost basis. And when businesses have a high fixed cost, the only way to survive is to operate at massive scale and distribute the fixed cost across many clients. Small, niche technology providers struggle to offer competitive pricing, and the barriers to entry become too costly for new entrants.
That consolidation has happened very quickly in the past year or two. But it’s a natural evolution for ad tech being integrated into these larger organizations.
