This interview was first in the Podcast Business Journal newsletter, with the latest podcast news and data. Subscribe free today.
Sean is CEO of Flightpath — this interview has been lightly edited for style and readability
Sean Howard: Flightpath is the predictive analytics platform for the podcasting space. If you were in the enterprise business world, you would know of SAS or IBM Watson or a few other predictive analytics tools, and there’s really nothing like that in the podcasting space for publishers that are looking to scale and want to identify opportunities to grow their business sell more inventory, leave less unsold inventory on the table, and free up their ad ops team to focus on improving customer service instead of putting out fires all the time.
James Cridland: So where did the idea come from? Was this an internal tool that you ended up building a while back?
SH: So I was a buyer in the space, buying podcast ads for podcasts. Working with Dan Misener, I was running Fable and Folly, which is an independent, one of the larger independent fiction publishers. I was building tools to help us scale the publishing side of the business with a tiny team. And on the buy side, it was frustrating how hard it was to give people money in this space. It becomes quickly apparent when you’re a buyer that people have no idea what they have to sell, and I’ve never been in an industry where that’s the case, where you don’t know what you have to sell or when. And it was frustrating to always be, as the buyer, to be the one finding every problem.
Flightpath was also for the publishing side. We were managing a lot of demands right from different buyers and from brands to DM across the board, and there were no tools to tell us what we had to sell or or to predict performance and delivery and to get us ahead of the curve - where we were, whether everything was going to run right. It just came out of that. And then it was actually Bryan Barletta who saw it one day. He called me and said “you got to show me this”. And when I showed it to him, he basically said “you’re an idiot, you’ve got to get this on the market”. So we started developing it for real.
JC: One of the tools that you have in Flight Path is flighting alerts, which tells you, in the middle of a campaign, whether or not you will actually deliver that campaign. What sort of shows work best with that?
SH: We have the first alert system that I’m aware of that works across order management systems and multiple DAI platforms - and it works at every stage. So we will alert teams to problems that haven’t even started yet. We started developing our algorithms working with fiction: which - as you know - there’s nothing “always on”. It’s all seasonal, and the dates are always shifting for launch.
There are nice things about an always-on show, but there’s challenges to that as well: how do you identify a sudden spike of of tire-kickers from actual, continuing listeners? So, our alert system is around all the things that can go wrong. If your sales team has sold a pixel, and there’s no pixel on the order three days out, we alert the ad ops team. If there’s a change in predicted performance in the future, we’ll alert the ops team. It just goes on and on. If DAI starts delivering over-cap, if you only want to deliver 200,000 impressions, some platforms will deliver half a million, but we’ll alert and stop that order.
JC: I’m guessing your tool must have spotted the changes in Apple Podcasts that happened towards the end of last year?
SH: For the longest time, Apple was saying “we’re not going to change anything”. Our tool was critical in being able to pace that down as it was happening. It was really neat watching how we could correct, on an hour by hour basis, the quotes going out from the sales team.
But it takes two to three to four weeks sometimes to close an IO. Our tools help you reassess, but you don’t want to go back and change things with the customer. Some people were really hit hard.
It was a horrifying experience for everyone. But from a math and data science point of view, it’s fascinating to see the shape of that curve and realize that it’s very similar to other black swan events that we have in our data set. That leads us to think - it would have been nice, going through it, to have been able to predict the exit. I guess the silver lining is, I think it’s now possible for us, in a future event, to add that. But yeah, it was a tough time for everyone.
JC: You work with a lot of podcast ad platforms and technologies. What have you learned about podcast advertising from it?
SH: I think, for those of us that are in podcasting, it feels old - but what I’ve learned is it’s still the Wild West, we are still figuring out how to scale businesses in this space. That is just new.
We saw a flurry of consolidations or shows being bought, and seen a lot of experiments. But I think the idea of a few people who met in their kitchen and started a show and pulled a few shows together and built a network… they’re now running multi-million dollar publishing businesses, and so those amazing content creators now looking to scale and grow from podcasting into other places is really neat.
JC: Are there mistakes that people are making that you see time and time again that your software can help with?
SH: Yeah. We all have to stop pacing on actuals. Our whole industry is built on “pace on actuals”. Every company is paying their teams - and it gets worse as they scale - to copy and paste data every day into Excel spreadsheets, trying to pace on actuals: you divide the length of the run, and you just divide by how much you know you’ve done versus how much you left, and then you try to hedge it. You’re trying to do all these things, but basically everyone is spending all their time monitoring spreadsheets and not the business.
I have seen some epic solutions out there, hodgepodged together at all levels. But I think when we start to see these companies invest in tools and software for the business, we start to see this amazing opportunity to free those people up to delivering a level of service that’s unmatched: actually being responsive, finding problems and contacting buyers months before their campaign starts to fix something. That’s unheard of in our industry, and it’s not in other industries.
But that idea of investing in the business and buying enterprise software is also alien to many of us in the space - myself included originally. At what point do we realize this is a really good business, and we can scale this, and what tools do we need to do that?
Something really amazing happens when we start to move away from “pace on actuals” into predictive, where your teams can start to be working months in advance and you can start to look at campaign performance rolled up in the future. You can actually start to look at your year in a predictive financial model - your full year of everything booked, and how you’re pacing in your projected financial performance this year. That’s a really fun change in mindset to watch clients take, and what that opens up for a business.
JC: Sean Howard, thank you so much for your time. I really appreciate it.
SH: Oh, anytime man Appreciate you having me on.
Get interviews like this every week, plus podcast business news and data. Subscribe to our newsletter free today.