Ensuring brand suitability solutions are suitable for podcasting.
![Ensuring brand suitability solutions are suitable for podcasting.](https://blog.tritondigital.com/hs-fs/hubfs/1720805839815.png?width=1066&height=511&name=1720805839815.png)
At the end of March, we announced our acquisition of Sounder - an AI-driven audio insights platform that transforms how publishers package podcast content and create contextual campaigns to align with advertisers' brand suitability needs.
Having worked with Sounder since 2023 in my podcasting role at Triton, I already had a pretty good handle on what brand safety was, could and should be for podcasting. But bringing Sounder into our fold, collaborating closely with their team, and exploring their tech up close has been nothing short of mind-blowing.
Here's some actual footage of me absorbing all of that information. Fun fact: this is the same technique I used in 2016 when I took over Omny Studio. Stick with what works, my friends.
My studying tactics aside, I wanted to share some of the learnings I've gained from overseeing Sounder and how it's reshaping my perspective on brand safety and suitability in the podcast industry.
Full stack or die trying
What was always clear (and a big reason behind the acquisition), is that unless a platform like Sounder is integrated at both CMS and ad server level, there would be numerous challenges for publishers wanting to package up their inventory in this way for direct sales.
Most shows see a majority of downloads happen in the first 24 hours after publishing. If you have to wait for a third party platform to ingest an rss feed before you can get the scoring back, you've already missed out on a decent amount of revenue. Public feeds also often have dynamically inserted ads, which will throw off the scoring.
We'd integrated Sounder into Omny Studio last year because we knew that we'd need the analysis of the episode back before publishing. Omny sends the episode to Sounder and waits to publish until the scoring is back and available for the ad server to read from. A matter of minutes makes all the difference here.
The ad system connection is even more important. For both direct sold and programmatically trafficked ads, the efficacy of a brand suitability platform is significantly diminished if the metadata isn't part of the targeting functionality of the ad server. More on this next.
Observation and adjustment vs targeting and curation
I'd had an inkling pre-acquisition, but what's become abundantly clear since integrating Sounder is that an observation only model in brand suitability does not scale for either the buy or sell side. Unless a platform like Sounder is integrated in the same way that geo-targeting or a DMP is built into the ad-stack, as well as offering a dashboard, risk and leakage remains.
To illustrate this point, let's look at two scenarios where a hypothetical agency or buyer wants to ensure ads run only on content that is marked as 'no risk' for Obscenity.
Scenario One: The buyer looks at a dashboard which provides a summary of brand safety scores for shows they've asked to be monitored by a provider. They pick the shows based on the dashboard and place the buy. Inevitably, they get busy with other work and only come back to check the same dashboard a week or two later. The show published a single episode that had an F-bomb in it a day after placing the buy. The additional time they were already going to have to commit to requesting the episode targeting is adjusted and verifying it all post-campaign, just got exponentially more complex. At their desk, they drop an F-bomb of their own.
Scenario Two: The brand safety system is integrated into the stack of the publisher they're placing the buy with. They send their Insertion Order or Deal ID through, along with the brand suitability requirements to hit. Because the scoring and targeting work in tandem, they know the ad will only run on episodes that match the criteria. It's no different to if they'd asked to ensure the ad targeted auto-intenders from DMP data like Adobe. Life is good, everyone high-fives and has time to go out for after work cocktails.
The above scenarios are obviously fairly simplified and dramatised to paint a picture (though, if you can't have a little fun when writing LinkedIn articles, when can you nowadays?) but my point remains - unless brand suitability or contextual targeting is integrated into the advertising stack, I don't think we'll get podcast buyers on and staying aboard this train this train.
Keywords aren't the key
Transcribing audio content is the first step when analysing podcasts for brand suitability or contextual targeting needs. But, as Sounder have known since first building their system, keywords in transcripts only get you so far when you're wanting to target to or away from certain topics. In the last three months, I've learnt that in a lot of cases, they take you down the wrong path completely.
Swearing is easy. As an Australian (not to cast aspersions on my fair country), I can 100% vouch for this fact. What I mean though, is there are only so many ways you can identify a swear word in a podcast and then assign risk based on industry standards.
Other brand safety categories get harder, fast. A podcast could talk about someone "dying from embarrassment" but it shouldn't be classified as a Harmful Act. A hockey player on a sports podcast could talk about all the great penalty shoot out's they've had, but it would be incorrect to assign this to Arms & Ammunition.
I thought this was all table stakes before we acquired Sounder, but I was wrong. And inventory being mis-scored is not good for revenue or podcasting in general as we navigate these new waters together.
Put this in context for me
To put the above another way, the context of the podcast audio is more important than keywords alone. Language is hard and words in one culture or dialect mean something completely different in another.
Believe me, as someone currently learning French and who has both accidentally implied their boss was less than chaste and asked a delivery driver to do something physically impossible to a door - context and culture make a difference. For those wondering, yes I'm still employed but no, I did not receive that FEDEX package.
As someone who feels very strongly about podcast revenues growing and more advertisers entering the space, I always appreciated that Sounder was aware of differences in language and culture. Their study with Urban One on diverse content, to make sure revenue wasn't being incorrect blocked, is something I share a lot.
With Podcast publishers and buyers in 86 different countries, we're focusing on training Sounder's models to retain this power as we roll out more languages.
Podcast publishers need to tell their own story
Not all podcast publishers want to introduce brand safety as a notion with their existing buyers - understandable. Advertisers who have already been buying shows from networks know what they're buying and why. Pure brand safety buys are not for these advertisers - they're for the thousands of brands who are yet to enter the podcast space (and likely won't until they can buy the same way as other media).
What's clear to me, is that even if podcasts don't want to sell this way today, that future is coming for us all. My hope is that by putting Sounder in their hands, they can better plan and prepare. I want podcast networks to be in control of their sales strategy and never be on the backfoot.
I'm encouraging our clients to think about brand safety as a way to avoid losing out on a campaign order, just because you can't check a box. Have that defensive card in your back pocket and start conversations around brand suitability with your advertisers while you're in the drivers seat.
Teamwork does, in fact, make the dreamwork
This is an obvious one and something I learnt early in my career - that a team makes all the difference.
When I came in as the CEO of Omny Studio, the development team was the first thing that stood out to me. Brilliant people, who are passionate about what they do and care about the end user.
I saw the same thing with the Sounder team and knew that post-acquisition I could expect the same experience that I got with Omny all those years ago. I haven't been disappointed and the product and development teams blow me away, every day.
Sounder wouldn't be what it is, without these talented folks. It's my privilege to work with them and help take what they built out to the podcast world.
That's a wrap. Whether you're a seasoned pro or a newcomer diving into brand suitability through podcasting, I hope you found this useful. Drop a comment below or feel free to reach out to me, if you'd like to chat about anything I've shared here. Thanks for reading.
Happy (and brand suitable) podcasting!