Is On-Demand Radio the Next Big Digital Channel?

The title of this post is really two questions:

• Is on-demand radio the next big channel?
• Is using “on-demand radio” instead of “podcast” in the title really just a sleezy linkbait move?

I’ll answer the second question first: “Maybe.”

Now, moving on to the first question. This post is in two parts: the first part is digital marketing prognostication, which isn’t tied directly to analytics. The second part is what the first part could mean for analysts.

The Second Life of Podcasts

No, I’m not referring to SecondLife (which, BTW, is still around and, apparently, still has life in it). I’m referring to the fact that podcasts just turned ten, and there are a lot of signs that they might be one of the “next big things” in digital. Earlier this year, when I wrote a post announcing the launch of the Digital Analytics Power Hour podcast, I listed three examples as to how  it seemed like podcasts were making a comeback:

• Slate published an article on how, while podcasting is a wildly imperfect technology, its 10-year birthday was a good point to reflect on the promise of the medium
• Serial took the world by storm, which prompted Fortune to muse on the growth of the format
• Alex Blumberg started a for-profit company — Gimlet Media — whose sole product will be podcasts (with the excellent Startup and Reply All podcasts both up and running), and he quickly hit his initial funding goal

The fact that I felt like podcasts were experiencing a resurgence should have been a death knell for the medium — predicting the future of technology has never been my forte (I distinctly remember proclaiming with certainty that SaaS-based CRM would never work when Salesforce.com launched! I said the same thing about SaaS-based web analytics!).

But, this past week brought another bullet for my list: The Slate Group launched Panoply, a podcast network where they’re partnering with Big Names In Media (think Huffington Post, The New York Times Magazine, HBO, Popular Science…) to produce high quality podcasts. Slate feels like they bring their experience (producers) and the technology (studio setups, audio mixing chops)to produce high quality podcasts. Many of the organizations who are joining the platform have a strong background in journalism, print, and digital publishing, but don’t necessarily have podcasting expertise. This seems like big news.

Here’s my list of why it makes sense to me that professionally-produced podcasts are poised for dramatic growth:

• On-demand TV has definitely gone mainstream with Hulu, Netflix, and Amazon Instant Video — while it’s odd to think that “radio” would lag behind “television” when it comes to innovation, TV seemed to be what had more advertiser focus (I don’t have a source to back that up), and the fact that it was video reasonably made it more attractive to startups/innovators; but, now that consumers are watching their television shows when and where (not just on a TV!) they want, it seems like they’re primed to forego the commercial-laden, live stream of traditional radio
• According to Andy Bowers (the long-time executive producer of all of Slate’s podcasts, and the main guy overseeing Panoply), the cost-per-user to reach a podcast listener with an ad is higher than even a Super Bowl ad; admittedly, it’s still a much smaller audience, but advertisers are starting to go in for podcasts, so media producers are picking up on that and looking to fill the space with advertising-worthy content
  
• Why are advertisers keen on the space? For now, at least, it’s a fairly captive audience. I’ll be the first to admit that I hit the “jump 15/30 seconds ahead” button when I hit the ads on many of the podcasts I listen to (although I also happily have paid to be a Slate+ member to avoid the ads on their podcasts), it’s still hard to miss the repetition of messages there. I wound up as a Carbonite subscriber for several years, and I became aware of them solely through podcast advertising. MailChimp is a perpetual advertiser, and these days, Cone, the “thinking music player,” seems to be betting on the medium.
• Not only are podcast listeners a reasonably captive audience, they can generally be reached on a consistent basis (like appointment TV viewing of days past), and, I suspect, are starting to fall into a pretty dreamy demographic for a lot of advertisers
• What’s in it for consumers? We’re getting more and more accustomed to a constant bombardment by media that has been algorithm-tailored or explicitly controlled by us to deliver information we want to see or read. Yet, when we get into the car or step out to mow the lawn (or shovel the driveway!), our vision is off-limits as a sense. The radio is the easiest stimulus to access that is an auditory-only experience…and that experience can often be pretty lousy. Podcasts give the consumer a much more self-tailored experience.

I get it. Being a podcast junkie myself, I have some inherent bias. But, the launch of Gimlet Media and Panoply has me thinking that it’s not just me.

So, What This Could Mean for Analysts

This, unfortunately, is going to be brief…and not particularly optimistic. The kicker with podcasts is that they are, fundamentally, powered with pretty crude technology:

• The podcaster creates an audio file and is responsible for hosting it somewhere (iTunes does not actually host podcasts — they maintain a library of pointers to the podcast audio files that are hosted wherever the podcaster has put them; Soundcloud actually hosts and delivers podcasts…but I agree with this skeptic about Soundcloud as a good platform for that)
• The podcaster updates an RSS feed — that’s right…RSS — that includes some basic meta data about the episode, including a link to the audio file
• Podcast hosting services get that updated feed and pass it through to the podcast platforms that their users use (the iTunes podcast app, Stitcher, TuneIn, etc.)
• The user’s local device gets that updated feed and downloads the audio file
• The user may or may not ever listen to the file, but it’s just an audio file — there is no universal mechanism to provide any data back to the podcaster about the number of times the podcast was actually listened to, much less how much of the podcast was listened to
• There is no universal concept of a “subscriber.” That’s something that is actually managed by the podcasting application. As such, any count of “subscribers” is, inherently, just an estimate

This explanation is just a way of saying, “This is why most podcasting advertisers harken back to direct mail by providing an ‘enter the offer code XYZ when you go to the site to get a special discount’ call-to-action.” There simply is not good data available as to the actual reach of any given podcast at any point in time.

I’m sure this is part of the reason that, occasionally, someone proposes that a fundamentally new technology should be introduced for podcasts. So far, though, that hasn’t happened. Yet, publishers are charging ahead with new content, anyway.

As analysts, we’re going to stuck with, essentially, three imperfect measurement tools:

• Download-counting — this isn’t bad, but it’s analogous to counting “hits” on the web (or, really, page views — it’s not quite as bad as the days of “hits” of any and all assets on a web page); it gives us a general sense of scale of the reach of the content, and it can provide some basic level of “who” based on the details included in the header of the request for the content.
• Offer redemptions — as I noted above, if the advertiser has a good CTA they can use where the listener will be incentivized to implicitly tell the advertiser “what prompted me to buy,” then there is a pretty strong link from a podcast listen to a purchase; unfortunately, the number of organizations where there would make any sense, while large, is by no means universal.
• Asking — several podcasts I listen to periodically embed a CTA in an occasional episode to go fill out a survey. I’m a fan of the voice of the customer, so I’m totally cool with that. But it requires reaching a sufficient volume of listeners to where a meaningful set of data can be collected that way. And, of course, there is a major self-selection bias risk — the listeners who are most likely to take the time to go to a link and fill out a survey are their most loyal listeners, rather than a representative sample of all of their listeners.

I predict this will be an interesting space to watch over the next few years. As a rabid podcast listener, I’m excited about the possibilities for the medium. As an analyst, I’m afraid I’ll be feeling deja vu when marketers first get serious about wanting to measure the impact of their investments in the medium.

What do you think?

Photo Credit: Patrick Breitenbach (Flickr)