Walton Isaacson’s Albert Thompson is speaking at the Convergent TV World conference on March 5-6 in New York City. Click here to register.
While marketers worry about AI disrupting TV ad buying, AI also has the potential to help break down the silos plaguing modern-day video advertising.
Despite industry discussions around channel convergence, “most marketers’ playbooks are actually not that diversified,” said Albert Thompson, director of digital innovation at Walton Isaacson. “They have the appearance that they’re diversified,” Thompson said, “but, in reality, there’s no intersection between their heads of investment.” Instead, they deal discretely with agency teams that handle broadcast TV, streaming, social video, etc.
Many marketers buy what’s familiar or “what they think makes sense,” Thompson said, partly because they’re beholden to traditional agency holding company structures.
But AI creates new opportunities for marketers to expand beyond their comfort zones. In addition to making video advertising more accessible, AI helps surface insights marketers need to make decisions about their investments.
Now that the video ecosystem is converging, buying workflows need to match how consumers actually watch content, Thompson said. “The number one trend this year will be gaining on AI.” (“Big surprise,” he added.)
I caught up with Thompson for a deep dive into how AI is changing today’s TV ad landscape.
AdExchanger: Could you describe the asynchrony you’re seeing between viewing habits and media buying workflows?
ALBERT THOMPSON: Fundamentally, content discovery on TV does not match how viewers actually look for content. Viewers search for content based on their mood, whereas streaming services and distributors package their content by genre. For example, a viewer might be searching for a show that’s “intense” and receive recommendations for horror movies when they were actually looking for, say, an action movie or a rom-com. The industry has to address this discrepancy to remove points of friction within the user experience, which should create more value for both sides.
There’s also a disconnect behind the scenes. Media buying structures remain massively siloed.
Agencies are used to earmarking media budgets by channel, such as social video or connected TV. But even within channel type, there are different buying teams. For example, TV or streaming investments are often split into different teams: one that deals in traditional network-based buying and another that deals in programmatic activations. Those silos mean brands accessing a broadcaster’s inventory can’t necessarily overlay programmatic audience targeting onto that activation, because a separate team handles programmatic.
In other words, a marketer could be dealing with three to four teams just to buy a video ad, which is ridiculous. Situations like these shine a light on the tug-of-war happening between marketers and agencies over the ideal way to activate a media campaign. A variety of investment teams arguably helps agencies boost their bottom line, but also creates a situation where marketers have to overspend on individual channels, which makes them likelier to stick to the channels and workflows that they’re most comfortable with.
How are marketers responding to this degree of fragmentation in the actual media buying process?
Media mix modeling (MMM) is playing a more significant role in media planning. MMM gives marketers a better understanding of how their media investments impact business outcomes based on their desired KPIs.
The insights marketers glean from MMM may justify reallocating media spend for better results, whereas current siloed operations fail to make that reallocation compelling. Marketers need a comprehensive view of their investments to optimize them for outcomes. In that sense, AI is highlighting the need to dismantle silos on the buy side so marketers can show up for their audiences in a way that’s more efficient – and drive more of the outcomes they’re seeking.
Coming full circle, how are you seeing AI impacting media planning and buying?
AI is making video production affordable for businesses without the budgets to buy long-form TV commercials. Cost effectiveness also means marketers can also make multiple creative variations for localized spots to help boost performance.
AI also helps marketers access video ad inventory and activate campaigns in a simpler, more cost-efficient way. That’s why we’re seeing startups like MNTN and Vibe.co crop up with self-service dashboards that emulate digital performance marketing. For similar reasons, content owners and distributors are also developing their own self-serve platforms. Comcast’s Universal Ads is an example, as are Roku’s and Paramount’s self-serve offerings. That development particularly makes sense for broadcasters with a broad taxonomy of historical TV viewing and programming data.
Content owners are also using AI for new features or visual overlays to bolster viewer engagement, including score tickers and game trackers appearing in live sports streaming.
How will marketers make use of new data and insights surfaced by AI?
What marketers really care about is audience targeting.
AI creates lots of opportunity for experimentation, including testing audience personas and proxies. The average brand is running creative against a series of personas or audience types that have unique overlap. AI cuts production costs so brands can test different creative iterations against different personas without flushing away half a billion dollars. Over time, those tests will produce insights marketers can use to ascertain the right channel, creative and targeting mix that drives desired outcomes.
Either way, the next disruptor of TV ad buying will be AI.
This interview has been edited and condensed.
For more articles featuring Albert Thompson, click here.
