Publishers aren’t selling what happened. They’re selling what’s going to happen.

Selling what hasn’t happened yet

Howard starts where most people in podcasting start, with what has not changed: it “remains one of the most intimate forms of media and continues to disrupt the traditional media industry.” Then the three things that have. The shift to video. New platforms, “podcasts are now on Netflix of all places”. And the spread of dynamic ad insertion, which is what turns an episode from a fixed object into inventory that can be sold and re-sold.

Underneath those sits a pressure that is less fun to talk about: how a podcast company or a creator builds a business that can survive changes in technology, platforms and buyer preferences, none of which they control.

What buyers now ask for follows from that. “We see a growing understanding of the need for a simulcast advertising strategy and platforms that can support and empower their teams to sell and forecast across media and platforms.” Selling audio and video as separate businesses has stopped making sense to the people writing the cheques.

And one thing they have stopped asking for, which is a surprising thing to hear from a company whose product reports on things. “We’ve moved past dashboards, partly because every publisher can use AI and other tools to create customized dashboards and other tools specific to their workflows and business needs.” The chart is no longer the product. Anyone can make a chart now.

Which leaves measurement, the thing podcasting has been about to solve for a decade. Howard declines the premise. “Media will always have measurement battles. Even TV and Radio continue to have advancements and changes to what they measure and how they measure it.” Whichever standard wins, what a publisher actually needs is to know that the platform their business runs on will represent their interests rather than someone else’s. And in any case he thinks the whole argument is aimed at the wrong half of the problem: “measuring what has happened is just the tip of the iceberg.” And then the sentence the rest of the interview keeps proving: “Publishers and creator networks aren’t selling what happened, they are selling what is going to happen, and that requires a level of forecasting and prediction that most, if not all, platforms struggle to provide.”

An early adopter of the unfashionable kind

The site says “prediction engine”, not “AI”, and we asked whether that was deliberate. It was, and the answer is more precise than a positioning choice. Flightpath is a traffic management system built on what he calls the first multi-platform prediction and forecasting engine in audio and video podcasting, and he will go one further and say it is the most accurate. The platform does use AI and machine learning. But the core of it “is giving sales and operations the tools to manage and scale their business, with a focus on maximizing future revenue. Something we can do better because of the forecasting engine we have been refining for over 5 years.”

That head start is machine learning, not the current wave: “one of the earlier forms of AI that came before the large language models we now equate with AI.” Being early to the unfashionable version turned out to matter.

What the fashionable version changed is pace, and then something less comfortable than pace. “The emergence of large language models has definitely increased the pace of feature development and it threatens just about any moat that is based on a feature set that is easily replicable.” It is an unusually blunt thing for a founder to say out loud about his own category, and it explains why he keeps pointing at the forecasting engine rather than the feature list.

The demand side moved the same way. “We see more clients looking for API and MCP access to our platform. Clients aren’t looking for dashboards, they are looking for data and insights they can’t get anywhere else.” And they need it across a growing number of platforms, because that is where the creators they support have ended up.

As for AI-generated shows turning up in the inventory he forecasts: yes, a few of the publishers and networks Flightpath supports have begun experimenting with AI content generation. “But it is far from mainstream. Podcasting, unlike radio and TV, loses a lot when that intimacy is broken.” The format’s defining quality turns out to be the one that is hardest to synthesise.

A system nobody understood, traded for an opinionated one

Work arrives through several channels, but the unusual one is deliberate: Flightpath runs what Howard describes as one of the few white glove professional services groups in the industry. “So much of the requests come from our team working together with our clients to understand and address any emerging issues or problems.” The roadmap is partly a by-product of sitting inside the customer’s week.

They started in YouTrack, for a sensible reason: the team uses JetBrains IDEs across departments. It did not hold. Tickets went missing, and the flexibility turned out to be the problem rather than the feature. “YouTrack is infinitely customizable but that also means it’s easy to end up with a system no one understands.”

So they went looking for the opposite, “a more opinionated ticketing system that would help to give our small team structure.” The thing that got their attention was not a feature. “Linear opened with their manifesto and we were intrigued. Moving to Linear has been a game changer for us.” The Slack integration and an early MCP server made the decision, in his words, a no brainer.

Those hours do not just sit in a report. “Time tracking is definitely taken into account in our annual pricing audits and reviews.” And the number he watches far more closely than he did two years ago is price-volume analysis, for a reason that points back at his customers rather than at his margin: to keep the pricing fair for smaller publishers while still accounting for the volume of transactions the platform processes.

Hire the ones who go looking for the pain

The company started as three people and is now ten, with the majority remote. Small enough that every hire changes the shape of it.

Juniors still get hired, through an incubator Flightpath was part of and still leverages for summer students, which has produced more than one permanent hire. What he screens for is not what a CV is built to show. “It’s not about how much you learn in books or how well you do on tests. We always look for the people who find a pain point not yet addressed and start exploring solutions.” After that it is whether they can work with the rest of the team, take feedback and contribute. And then the tell: “those who end up owning a new project or feature or chunk of code are hard for us to not hire.”

Asked what someone starting out should learn today, he gives the least technological answer in the interview. “How to build things and work with a team. If you can’t take feedback, that’s going to be a problem.”

What AI changed is the floor rather than the ceiling. “Someone who is reluctant to use an LLM is going to find it difficult to get a job as a coder.” He does not stop at engineering either: there is still room in the creative roles, “but that’s not the case in finance, strategy, operations and more.”

And what is left for humans in five years is, in his telling, the same thing he hires juniors for. “Identifying unmet needs and pain points (opportunities) and then finding creative ways to deliver value and solve these.”

Four short ones

One podcasting cliché you don’t believe any more?

Podcasting is as defined by the audience. YouTube is podcasting just as much as listening to it in an audio podcasting app.

Which piece of common advice in your industry is wrong?

There is a lot of confusion around what a simulcast sales strategy is, and it is critical for growth and success in this industry.

Something you would never automate?

Customer support and success.

A prediction two years out you’d bet money on?

More consolidation, and a growing awareness among publishers that the platforms they choose to be on may not always operate in their best interests.