Skip to main content
Spotify Erroneously Serves Liquor + 4 more stories for May 27, 2021
"

Spotify Erroneously Serves Liquor + 4 more stories for May 27, 2021

The Download

Season 0 • Episode 23

This week on The Download: Spotify erroneously serves liquor, platforms continue to snap up content houses, and RSS.com shares how they beat podcast spam.

Last week Spotify experienced the mother of all brand safety mistakes, and for now it has cost them the ad dollars of a major liquor distributor. As originally reported by Tanner Cambell in Saturday’s issue of Dispatches from the Booth, sometime on Saturday Spotify began serving a banner ad for bourbon distillery Wild Turkey to the profile of every podcast. This led to numerous screenshots demonstrating just how unfortunate such a slip-up can be, including children’s entertainment and podcasts discussing sobriety. On Tuesday Wild Turkey made an official statement on the matter. Quoting their statement:

“We are disappointed that the implementation of the campaign by Spotify was inconsistent with our advertising code and have paused all media with Spotify pending results of the investigation. We are members and partners of DISCUS and Responsibility.org and unequivocally support a responsible marketing agenda.” 

The Download has covered numerous stories involving new tech designed to ensure brand safety when it comes to problematic podcast hosts or suspect content in an existing brand’s catalog, but this slip-up shows brand safety doesn’t stop at the content of the podcast itself.

Which leads to a decision that might set up further problems in the very near future. This Monday Issie Lapowsky published an article for Protocol highlighting Spotify lifting its two-year ban on political ads. Quoting the article:

“In an email the company sent out to potential partners this week, Spotify said that political ads will appear “across thousands of podcasts on and off Spotify.” An accompanying presentation promises political advertisers the ability to target niche audiences and tap into AI-driven “contextual targeting,” which allows advertisers to place ads in podcasts when they are discussing issues relevant to their target audiences.”

Lapowsky highlights the uphill struggle Spotify has ahead of them. While competitors like Meta and Google have made efforts to be open with political advertisements by maintaining large public archives, Spotify’s starting from scratch.

The Wild Turkey slipup isn’t enough to raise alarms at Spotify’s ability to run ads ethically. Mistakes happen. That said, it does highlight just how impactful a mistake can be when made by a massive player in the industry. A player now dabbling in an area of advertising infamous for its ability to spread disinformation.

It’s been a week of big-name acquisitions of talent in the podcasting world. The massive long-running interview podcast WTF with Marc Maron signed with Acast, according to a Brad Hill article published Tuesday. 

“‘Entering into a partnership with the podcast giant Marc Maron is a big win for Acast and for the open podcasting landscape, as this deal guarantees that the podcast can continue to be listened to by everyone, everywhere and on their preferred listening platform,’ said Ross Adams, CEO of Acast.”

WTF, which previously operated as an independent production with a paywalled back catalog available for a fee on its website, will now switch to offering older episodes to Acast+ subscribers as bonus content. For the audience little will change, but Acast has just inherited one of the biggest names in podcasting outside of Spotify’s walled garden and gets to handle ad sales.

Establishing the trend for the week, Monday saw a press release from SiriusXM announcing the acquisition of Team Coco, most notably bringing aboard Conan O’Brien Needs a Friend. Like the Maron buyout, Sirius has purchased the production house whole hog, including a five year talent agreement with O’Brien. Longtime fans of the former late-night comedian will understand the gravitas behind convincing O’Brien into an exclusivity deal after his messy breakup with NBC.

Rounding out the trio of announcements, Monday also saw RedCircle’s announcement of acquiring rights to travel influencer Drew Binsky’s upcoming podcast Roots of Humanity. Binsky is making the popular move for influencers by starting a podcast guaranteed to court a built-in audience from his existing audience, sporting 3.3 million YouTube subscribers alone.

“Roots of Humanity celebrates the beauty and diversity of the world, which is aligned with RedCircle’s vision of helping podcasters of all sizes and shapes across the globe to get rewarded for their work. RedCircle is excited to be part of Drew Binsky’s journey and help him to achieve his goals with its modern podcast hosting and monetizing technology.”

It seems the trend continues with podcast platforms looking beyond simple exclusivity rights for an individual podcast to focus on acquiring production houses and existing brands that can act as content engines beyond the popular podcast’s lifespan. Full creator/talent packages are the current hotness and there’s no signs of them slowing down.

This Monday Sounds Profitable’s own Bryan Barletta covered the launch of Buzzsprout’s new ad project in a thread on Twitter. In simple terms, the Buzzsprout Ads beta gives podcasters the ability to use Buzzsprout’s dynamic ad insertion tech to implement midroll ads in a way previously only available to signed talent or shows with crew who’ve worked in advertising.

Producers entering the beta can manually review ads to decide which would be the best fit for their podcast and whitelist them for inclusion. Buzzsprout’s tech by default uses a music jingle slowly increasing in volume to signal an ad break is approaching, attempting to emulate the five second timer announcing ad breaks on video hosting platforms like YouTube. Users are given a surprising amount of customization with this feature, offering five stock transition options as well as the ability for hosts to record their own ad break and ad return bumpers for the automated system to use.

Accessibility has also been taken into account. The Buzzsprout Ads platform is designed to drop chapter markers around each inserted ad and automatically time-shift the SRT file for podcasts with transcripts to accommodate the inserted ad changing timecodes for any portion of transcript that takes place after.

Bryan sums things up best at the end of his thread, saying:

“Ad monetization empowers many podcasters to grow their show into a business. Sponsorships will always take the cake, but out-the-gate options for a skilled creator without sales experience changes the game. DAI Marketplaces and Programmatic are how we hit those IAB numbers.”

A few days ago a string of spam podcasts were published to major platforms, all containing a single episode of unrelated audio aimed at using the show art and description fields to promote escort services in various locations in the United Arab Emirates. They had the legitimacy of the spam texts one gets at 2:00 a.m., but they were produced by the dozen. On Monday RSS.com co-founder and managing director Alberto Betella published a writeup on Podnews detailing how the hosting service tackled their new spam problem. 

“We had two options: a. “Throw humans at the problem”, hire 1 or 2 extra people to counteract this misbehavior (perhaps also implementing a moderation queue for newly created podcasts). b. build tech to address and mitigate the problem.

We chose option B because it works at scale, it doesn’t need food and drinks, and because it’s a lot more fun! A few people in our team, in fact, have a strong AI/Machine Learning background and it doesn’t happen very often to have the opportunity to leverage this skill set in the podcasting space.”

What’s important here isn’t necessarily the fact RSS.com was able to quickly fix the issue with neart machine learning and clever implementation of tools to prevent false positives from getting stuck in the system, it’s the fact that we the public know the broad strokes of how they accomplished it.

“We cannot share our current ML model as open source because it is part of the unique value proposition of our company. However, we have described the methods and provided actionable recommendations for anyone that wants to build something similar. Hopefully, this is our small contribution to keep podcasting spam free!”

Sharing methodology used to overcome issues that could affect anyone with similar services in the market is useful, embodying the spirit of open source communities even when dealing with proprietary creations. We need more sharing like this across the industry.

Finally, it’s time for our semi-regular roundup of articles that didn’t make it into today’s episode, but are still worth working into your weekend reading.