Erik Diehn: There’s no silver bullet. There are people who have ‘unfair’ advantages in that, for instance, they are a show produced by NPR and they have vast national radio distribution and ten years of audience history.

For the pure podcasters or new shows, having an audience that already exists, either for your brand or for you as a talent, helps, but is neither a guarantee of nor a requirement for success. We have a couple of YouTube stars like Grace Helbig and Shane Dawson who started out on YouTube and grew large audiences there.

Certainly [you need] consistency of output; you need to pay attention to the details. If you have a podcast with a difficult title and terrible artwork, it’s a strike against you. A lot of people overlook those finer details, but in a crowded marketplace they can make a whole lot of difference.

And then I think quality and passion and hard work — again, not a guarantee, but in general, people who put more effort into their work tend to get more of a return. Comedy Bang Bang may seem like a freewheeling improvisational show, but Scott Aukerman actually puts a great deal of thought into the structure, the guests, the cadence, the rhythm, the final product. WTF is Marc Maron‘s full-time job.

You need a distinctive voice. An awful lot of shows come to us that have celebrity guests, and they’re surprised that their numbers are not huge, just a few thousand. The reality is, people may sample you for your guest, but they’re only going to subscribe if they like the show. People listen to Fresh Air because they want to hear Terry Gross, even though she may only be doing 15 percent of the talking. You have to have that distinctive voice; you have to have something that cuts through.

Diehn: I can’t divulge revenue numbers, but I can tell you in general that the top five shows are probably pretty close to what you’d expect, and they correlate with audience size. There are a couple of surprises that you might find, but WTF is a significant earner, Comedy Bang Bang, How Did This Get Made, Nerdist.

Then there are some surprise ones. One I can talk a little more publicly about is this show called EOFire — that’s a show with an audience of 60,000, maybe it’s a little bit more now. But [John Lee Dumas] publishes daily and he has an exceptionally engaged audience, and as a result we get a lot of advertiser interest and a lot of advertiser renewal. He is consistently a top earner in our portfolio, even though his show is much smaller than, say, WTF.

Diehn: To some degree. There’s a scale problem, though: If you have, for example, a pet-focused podcast, it’s not as if there are twenty pet-focused consumer goods companies waiting in the wings. We’re just hitting a point where we’re getting larger brands buying, and every once in a while you’ll find that kind of niche connection.

It’s hard to justify going out and mining the waters of, for example, home brewing, where there’s a limited number of advertisers trying to reach an audience that very much wants their products. The matching game is hard to do at the scale we’re at today. If we get critical mass and suddenly we have five craft brewing podcasts, then it makes sense to go out and find ten craft-brewing supplying companies and match them up.

We had a show called Cardboard that Rich Sommer from Mad Men did for a while with us (it’s on hiatus now). He just loves board games. He played them on the show, he talked about them. Even though we only had an audience of about 10,000 to start (and that was growing because it was a really good show), it was constantly sold out. Board game manufacturers knew that was the exact perfect place to advertise. That’s an example of how that can work. It’s hard to get a lot of these examples, though.

Diehn: It is. Everyone would love to have YouTube-style metrics. But the measurement of podcasting is really not that much worse than any other medium; we just don’t have a single source that everyone endows with some sort of holy status.

Arbitron, which was the measurement standard for radio for years, was based for a long time on people writing down on a piece of paper what radio station they listened to that day. The sample size was literally thousands of people, for all of New York City. Literally billions of dollars changed hands based on less than a hundred thousand people writing down on a piece of paper what radio station they listen to.

That got better when Arbitron gave those people a little pager device that listens to an inaudible tone that radio stations broadcasts along with their signal, but then suddenly those people were picking up things that were just playing in the bar they were walking past, and they had to remember to carry the pager with them. There were so many problems with that system, but again, even more billions of dollars transacted based on those measurements.

So when people say, ‘We don’t know how many people listened all the way through to the last minute of this podcast,’ my feeling is, yeah, but we know a lot more than we did 20 years ago, when radio was based on the whims of a handful of people. We’re certainly getting better. And some of the best measurement is actually through our advertisers. Today, direct-response advertisers are not trolling late-night cable trying to get people to dial the 1-800 number. It’s sophisticated, direct-to-consumer, digital, e-commerce retailer subscription services.

Diehn: We wanted to create alternative revenue streams so that we’re not entirely dependent on advertising — to some degree, taking a page from the playbook of public radio or cable television. Also, it can be hard to do really interesting audio things that are limited series, short-run, that are kind of weird and quirky, in a regular ad-supported weekly or biweekly format. Many times over the years, we would talk to talent who would say, ‘I would love to do something with you guys, I just can’t get into a weekly podcast.’ Which stinks!

Our view is that if we find an audience that loves this type of content, eventually we can start expanding out from there. We can get a recurring subscription from those people that will create a pool of revenue that pays for this content to be created. This is exactly what Netflix does, it’s exactly what Amazon does with Prime. It’s an ecosystem where the consumers are directly funding the creation of that content.

We can do interesting, offbeat six-part series. We can do highly produced things with Jemaine Clement from Flight of the Conchords. We can have relatively up-and-coming comedians do these insane parodies of Dan Carlin’s Hardcore History that are all about Game of Thrones — that show happens to be one of our highest performers in the past couple of weeks. We never would have taken a risk on that kind of show in an ad-supported environment.

Diehn: I don’t think we’ll have a sudden hockey-stick moment where we go from 17 percent of the U.S. population [listening to podcasts] to, say, 80 percent. I think there’s going to be the same upward trajectory for the next few years as demographics continue to move in our favor, and as people listen to their phones in their cars more. It will just get easier.

Diehn: There’s a transaction cost associated with podcasts for every listener. It’s a commitment, a choice. You don’t do it for something you just don’t care that much about.

Radio is a passive medium that has almost no transactional costs. You get in your car, you hit the power button, and if your radio is already pretuned to the station and your volume is set right, you do nothing until you arrive at your destination. You’ll just take whatever streams through, and as long as that content is of a level of quality that is reasonably satisfying while you are driving your car, you’re pretty happy with it. If it’s better than that, maybe you’ll donate to your station.

Those three- to five-minute news segments strung together can be high quality, but people are not going to go, ‘I want to listen to this three-minute news segment now,’ because transaction costs associated with getting that news segment are so high relative to the value you get out of it.

That’s why program directors have multiple things on throughout the day. They have Morning Edition, then Brian Lehrer, then The Takeaway. You’re programming around the clock. Podcast listeners have their own clock, and they program their own experience.

Public radio might need to ask, ‘Are we all about producing Wait Wait, Don’t Tell Me? Or are we all about taking grant money and other things to dig deep and turn up stories in our local community that nobody else is going to turn up?’

Public radio is pretty good at innovating, but I think that in order to survive, they need to figure out how to make the commercially interesting stuff live alongside the mission-driven stuff, and have one fund the other without stepping on each other’s toes.

Diehn: It would be nice to have consistent client-side measurement. Even in other media you don’t necessarily have that, though. We’re closer in web video, in TV. But we’re pretty far from that in podcasting.

It would be nice if we got to a scale where we had a little more infrastructure. We don’t have enough scale to track tons and tons of venture dollars coming in, building ad-tech platforms for audio like we did for video, at least not for podcast audio.

It stinks that there’s not better stuff out there. There are a few things emerging, but we’re still frequently in a situation where we’re like, ‘Do we have to build this? We don’t want to be a tech company! We’re a media company.’

I’d also like to see the audiences grow and diversify. I think they already have, a lot. We’re getting a broadening of content and audience, but it needs to continue to grow. If I could just wave the wand, I would want to have a broader cross-section of the public listening tomorrow, and a larger percentage of the public listening.