AI and identity: Unlocking the future of podcast ad targeting

If you’re like me and the luxury of screen time eludes you, there may be no better medium for mobile learning and entertainment than podcasting.

For every audience, regardless of how niche their interests may be, there are virtually limitless and, for the most part, free shows just waiting to be discovered.

Many marketers in our space revere podcasting and embrace it as a part of their daily lives, but the brands they manage are leaving podcasting off media plans.

How can podcasts attract more investment from advertisers? As an admittedly biased member of the digital audio industry myself, I have a few ideas. But first let’s get to the facts.

Podcasting growth and investment

Edison Research reports that the percent of Americans ages 12 and older who listened to podcasts has grown steadily over the past two decades, surpassing the 50% mark in 2019.

Pew Research Center reports that even though their content preferences may vary, U.S. podcast listenership per capita is roughly the same across many racial and ethnic groups such as Hispanics and African Americans.

Triton Podcast Metrics Demos+ reports show a 130, and 128 index for college graduates and Income $100K+ demographics respectively, according to their 2023 survey of 24,000+ monthly U.S. podcast listeners.

Despite this irrefutable data that podcast listenership is growing amongst a diverse, well educated, and higher-income audience, digital audio still generates far less ad spending per hour listened. In 2024, digital audio — because the eMarketer report does not break out podcasting versus streaming music — was forecasted to generate only $36.1 per hour listened, paling in comparison to the likes of digital video and social media, which generate a stunning $504.5 and $420.3 per hour of engagement.

If podcasting is such a widely adopted medium, why is it escaping most media plans?

Podcast brand safety and brand suitability

Many of you are familiar with the legacy brand safety and brand suitability technologies that exist for display and video. Keyword inclusion and exclusion, along with pre-bid blocking solutions.

Audio, like CTV, has one major nuance that prevents media buyers from implementing the same display and video advertising safety tactics. Audio does not exist on a website. For a long time, this has plagued the programmatic audio community.

But now, AI may have a solution to ease advertiser concerns around brand safety. Companies like Sounder use AI to analyze transcriptions of podcast content so advertisers know their message is surrounded by brand safe content.

I am beyond excited about this advancement. Advertisers deserve to have a better solution to podcast brand safety than what existed during the years prior to AI scoring. Cautious media buyers will tell you they would need to manually comb through hundreds of different shows to make their case for an allowlist for their brand. It is hard to think of a better way to shrink a podcasting budget.

Another major ramification of list-based brand safety is the exclusion of “long tail” podcasts. The thousands of shows that don’t have the same listener volume as the Joe Rogan Experience, but still have meaningful listenership, are overlooked during manual review simply because of the extra work required to research them. Your brand’s favorite audience could be scattered amongst the medium to long tail of podcast shows, and you would never know it. With the power of AI solutions, we can finally give brands a chance to uncover these underrepresented, but worthy audiences.

These AI algorithms are even capable of contextualizing how words are used by the podcast hosts. No need to restrict your campaign’s reach by blocking the word “bomber” unnecessarily if, say, your audience happens to be listening to a baseball podcast that mentions the “Bronx Bombers” to refer to the Yankees. As a Mets fan, I’m not terribly worried about that scenario. As a podcast industry professional, context is everything, and we need to know what hosts are talking about before blocking their shows.

Podcasting’s identity crisis

If you’ve ever had firsthand experience with audio, this one comes as no surprise. Podcasting — because users do not interface with a web page before consumption — will not share the same signals that buyers tend to rely on to serve targeted ads. 99% of the time, you will get the IP-address of a user and not much else.

Podcasting is highly distributed, meaning that the platform of origin, let’s say for example iHeartMedia, may not be the application where listeners are tuning in. It may be on Spotify or Apple podcast platforms. Distribution means less information is being shared back with publishers about their listeners, and that translates to minimal targetable signals for buyers.

Unless we want listeners to shift over to reading their podcasts from a web page — as fun as that sounds it’s clearly not going to happen — buyers, publishers, and even podcasting platforms need to work together to tackle this one. Luckily, there’s a ton of creative options on the table.

One of my favorite tools to bolster addressability, and measurability of podcasting, is identity graphs. If a podcast impression can provide at least the IP address of a user, identity graphs will allow buyers to understand what kind of listener they can target.

Brands will never stop leaning in on audience targeting, they know their customers too well. If the supply side of the audio industry wants the revenue our channel deserves, we need to work with, not against, the buy side to make media useful to them. We don’t have the luxury of readily available identifiers, so it’s all about adopting many addressability solutions to test out what works. Triton’s Partnership with ID5, for example, leverages a variety of consented signals provided by media owners to resolve identity when some “walled gardens” won’t pass back user-level data. It’s all about getting creative and establishing the best possible relationship with the player applications as well as buying platforms.

Everyone loves a prediction

Podcasting is primed and ready for mainstream investment. I’m optimistic about our industry’s chances of seeing a major jump in investment in the next few years. Albeit increasingly fragmented, podcast platforms have been making the right moves to make this channel more accessible for the buying community.

This op-ed represents the views and opinions of the author and not of The Current, a division of The Trade Desk, or The Trade Desk. The appearance of the op-ed on The Current does not constitute an endorsement by The Current or The Trade Desk.