Podcast Advertising at Gimlet Media

Source: gimletmedia.com

A few days ago, as I was catching up on Reply All episodes I had missed over the last few months, a few questions kept popping into my head…

Why is every single ad a Squarespace ad? Squarespace is cool, but every single one? Really? Doesn’t Reply All have other sponsors? Did Squarespace eat their ads?

So, I listened to every advertisement of every episode of every show Gimlet has produced currently on iTunes and cataloged it. Here’s my spreadsheet, it looks a lot like this:

Source: My Computer

Why do that? Two reasons:

  1. I really appreciate how Gimlet approaches advertisement. At least two Startup episodes (But Wait, There’s More and We Made a Mistake) have dealt with this very issue. As podcasting continues to develop, Gimlet’s method of building empathy/understanding around advertising through its shows will be crucial.
  2. Gimlet can do better! They have an incredible platform for advertising, but in my opinion it is not being fully leveraged. But assessing its current status, I can better understand its potential.

After completing my analysis, here’s are my high level recommendations, with more detail below:

  1. Shift more ads from lower CPM post roll spots to higher CPM pre roll spots and use Gimlet’s trademark storytelling power to deliver a great ad experience for listeners that they wouldn’t want to skip.
  2. Use Megaphone to collect data that will better define user metrics and lead to efficient advertisement scale/distribution across shows.
  3. Diversify advertisers who buy non-premium ads (spots that are more than ~60 days old) to not be over-concentrated in an individual advertiser.

Analysis

  • Gimlet has produced 99 unique episodes with advertising content (this excludes 5 episodes that didn’t have ads). These 99 shows total 46 hours, 39 minutes and 17 seconds worth of airtime.
  • Within the 99 episodes, there are 243 ads that run a total of 2 hours, 47 minutes, and 46 seconds. For listeners, this means on average you get a little over 16 minutes of Gimlet content for every 1 minute of ads.

Those 243 ads make up Gimlet’s current ad inventory, which needs to be segmented further into pre roll, mid roll, and post roll ads.

  • On average, a pre roll ad is 27 seconds long, mid roll ads average 64 seconds, and post roll ads average 17 seconds.
  • Of the total 2 hours, 47 minutes, and 43 seconds of ad time, mid roll ads take a whopping 2 hours, 9 minutes and 43 seconds of total time.

This makes a lot of sense for both the listener and the advertiser. For listeners, they’ve already consumed about half the story and are more than willing to take a quick ad break. Advertisers understand this, so they load up the middle of a podcast with relatively longer ads to tell their own story at efficient cost.

That being said, advertisement at Gimlet could be better optimized by smoothing out the difference between post roll and pre roll ads. Without knowing the exact numbers, it’s safe to assume that pre roll ads command a higher CPM rate than post roll ads. Listeners are probably less likely to stick around at the end of an episode for the post roll ad vs. shutting off a podcast episode at the beginning because of an ad. By moving towards an ad distribution that Mystery Show currently employs (11 pre roll, 12 mid roll, 0 post roll), Gimlet can capture those higher rates and generate more revenue.

The counter-argument to this is the ability for listeners to skip ads entirely with a few taps of their finger on whatever podcasting app they use. This would obviously decrease the CPM rate of a pre roll spot. But, Gimlet’s advertising has two distinct advantages:

  1. When applied to ads, narrative storytelling makes them surprisingly enjoyable, and nobody does storytelling better than Gimlet. By having the ability to work directly with advertisers, conduct interviews for ads, and shape the creative direction, Gimlet can deliver a better experience for the listener and better results for the advertiser.
  2. The ads are read by the host(s) of the show. This is a voice the listener already trusts and expects to hear at the beginning an episode, which may decrease the chances of them skipping the ad.

Of course, this isn’t a perfect science yet. Many of these insights would be easier to validate with listener data to back them up. It’s likely that some optimization model exists to maximize the relationship between listener experience and all the variables that go into digital advertisement placement.

Podcast data is not easy

Think about what would go into that model. It would have to include listener data segmented by platform (iTunes, Soundcloud, Overcast, Stitcher, etc.), audience size (downloads, streams) per show and per episode, engagement (when do listeners stop listening to an episode, what time of day are they listening), and demographics, to name a few. It would ALSO have to include data on your advertisers: ad rates, length of average campaign, total number of campaigns to date, total current ads across shows (segmented by pre/mid/post roll).

The point is there’s a TON to keep track of as podcast companies like Gimlet reach for that optimal point. This is where Megaphone comes in.

The Megaphone Announcement — Source: Panoply

A few years ago, Panoply, the podcast arm of the Slate Group, bought an Australian company called Audiometric. As of recently, they are reintroducing it as Megaphone and have signed Gimlet as one of its first customers. The technology “features include centralized management of ads, podcast episodes, feeds, and distribution. It has real-time, dynamic ad insertion; performance analytics for ad campaigns; geotargeting and copy-splitting campaigns; and wide distribution of ad-supported podcasts.”

If effective, this is the kind of technology that leads to massive scalability. Gimlet will be able to make the right moves, quicker, with fewer resources expended and a better understanding of the growth opportunities in their shows, both on the revenue side and creative side of the business. It would also help eliminate mistakes and make substituting ads seamless. For example, it could identify the mysteriously unutilized ad space in Reply All episode #23 at 13:15. It could also eliminate the awkward gaps in audio like at the end of Reply All episode #22.

I couldn’t be more excited for Megaphone. This is what will make advertising so exciting at Gimlet moving forward — helping shape the tools that will not only grow a company, but impact an industry as well. Selfishly, access to Megaphone would’ve saved me a lot of time on this little project too. If I had access and was looking at Gimlet advertisers, I’d probably start with something like this.

That’s a lot of numbers. But there’s magic insight here, and it answers the question that started this whole project.

What’s Going on with Squarespace?

  • Of all the 168 Squarespace ads currently occupying Gimlet ad inventory, 97 of them can be found in Reply All episodes. That’s 57.7% of total inventory, and 80.8% of Reply All inventory.
  • As a show, Reply All only has four current pre roll ads, which means Squarespace dominates the mid roll and post roll spots. Yes, Reply All does have the most episodes of any Gimlet show, but they also have the most listed sponsors of any show on the Gimlet website.

This piece of information confused me. Where did all those other sponsors go!? The answer highlights something unique about podcast ads, and frankly all digital media that is not moment dependent (aka live).

These ads are not static, they can be swapped in and out of episodes as current episodes fade into history and new ones replace it. Ad spots on current shows would be considered a premium ad spot, unless Gimlet can prove they have a generally consistent monthly audience across all episodes of a show at a given point in time (another piece of data to look at!). But, how many times do you think your average podcast listener goes back to listen to a single, individual episode?

Squarespace lives in this ad space. Reply All has other sponsors who’s advertisement campaigns ended and were replaced by a Squarespace ad (which likely cost a little less). If I had to guess, I’d bet the Squarespace insertion order (IO) with Reply All outlines nearly exclusive rights to mid roll and post roll ads in past episodes prior to December 16, 2015. This also explains the discrepancy between the actual length of time of an episode and the stated length of time on the podcasting app. For example, the Reply All episode entitled “I Want To Break Free” was 25:39 long when originally posted, but now runs 28:46 as a result of the added Squarespace ads.

Gimlet is currently over-concentrated in Squarespace ads, which poses both a business risk and an opportunity cost. As the podcast content landscape develops in competitiveness and average show quality improves, the chances increase of Squarespace having access to a similar audience size at a lower cost than Gimlet is currently providing. This builds a case for Gimlet to mitigate some of that concentration risk by seeking different advertisers for those non-premium ad spots.

Additionally, this resale feature of podcast advertising has amazing potential because it opens up the digital advertising space to companies with a lower marketing budget. If every ad commanded a premium CPM, the market would effectively lock out a huge source of potential demand: companies that would love to advertise with Gimlet, but don’t have to resources to make the economics of a premium spot work. By developing relationships with those types of advertisers early and often, Gimlet can tap that potential demand which may lead to larger IO’s in the future.


Finally, there are two things I’m certain of.

  1. This analysis just scratches the surface. There are obviously many more conclusions you could draw from the data spreadsheet, and I hope it serves as a starting point for anyone else looking to dive in!
  2. This space is poised to change dramatically in next 12 to 24 months. What will be the effect of better technology on ad rates? Will ad rates suffer from increased interest from non-traditional advertisers (more demand)? Will more and more podcasts see monetization?

Gimlet, I’m sure, will be on the forefront of these conversations. Thanks for reading!


Further Reading:

  1. But First, A Word From 100 Podcasts’ Sponsors — FiveThirtyEight (5/1/15)
  2. Why So Many Podcasts Are Brought to You by Squarespace — The Atlantic (5/12/15)
  3. Anatomy Of A Podcast Ad — FiveThirtyEight (5/1/15)