I almost started a podcast agency. Here’s what happened.

During my senior year at Stanford, I squandered a big opportunity.

It’s a mistake I hope to never make again.

Here’s what happened.

The entrepreneurial fire

During my sophomore year, I had cofounded an influencer marketing agency, and it had really taken off.

We were making tens of thousands of dollars per month and left Stanford to pursue the opportunity.

Due to cofounder disagreements, I left the company and returned to campus. (side note: start companies with people you know)

Despite the lows, I had learned that you could just interview people, solve their problems, and collect their money.

I was hooked.

I began immediately looking for a new business idea to pursue.

And during my senior year, I found an incredibly good one.

Podcast advertising

I had become obsessed with the podcast advertising market.

It was 2015, and Serial had put podcasts on the map. They were just starting to go mainstream.

Podcast advertising - I thought - was cheap. Way underpriced.

My hypothesis was that this would be a huge industry — that some advertisers were already in on the secret and that they were CRUSHING it with their podcast ads.

Validating the niche

So I got to work.

I found a list of all the top podcast advertisers, and I made a spreadsheet of all their marketing executives.

I used a free email verifier to reverse-engineer their email addresses and started reaching out.

My goal? To validate my hypothesis.

To my surprise, an enormous percentage responded and were willing to hop on the phone. (about 15%)

In short, I was right. Podcast ads were KILLING it for them. The ROI was phenomenal, and they couldn’t get enough.

From my conversations, it felt like there was infinite demand for podcast ads. They were so damn profitable to run that advertisers would bend their budgets to spend more on them.

I asked these executives what problems they had, and there were three common answers:

  1. coordination was annoying (i.e. managing all of the conversations with podcasters)

  2. there wasn’t enough supply

  3. it was hard to track performance in a granular way, and ads were sometimes “stale”

Let’s dive into each of those.

Coordination was annoying

These brands were directly talking to various podcast networks and individual podcasts, and those conversations took up a lot of time. They wanted a way to “spend money, get ads”.

hint: this was a massive opportunity. We’ll come back to this.

There wasn’t enough supply

At first glance, this seems very self-explanatory: not enough stuff to spend money on.

But there’s more to it.

Since ad-buyers found coordination cumbersome, they were unwilling to work with podcasts that didn’t have a lot of listeners.

At the time (and today), most podcasts had small and medium audiences. (i.e. there’s a "long tail" of podcasts with small audiences)

This means that there was a lot of supply that you could "unlock" if you were willing to do the coordination.

Performance tracking & stale ads

The metrics around podcasts were a bit flimsy. Podcasts reported their download numbers, but it was impossible to know whether a user had actually listened to an episode that they downloaded.

Plus, if a listener heard an episode from a year ago, they’d hear an ad that was a year old, and that ad may not be relevant anymore.

Many advertisers mentioned that they would love more granularity on actual listenership numbers, and several mentioned that dynamically inserting ads at “listening time” would allow them to run way better campaigns.

But here’s the thing: in retrospect, I don’t think that was really a problem.

Yea, it might have been cool to get better metrics or just-in-time ad insertion.

But advertisers were still spending buckets of money on podcast ads even without these features.

My fatal mistake

As a coder, I got tunnel vision.

“Great,” I thought, “I’ll build a podcast app that dynamically inserts ads at listening time and provides granular analytics to advertisers.”

The problem?

This was an enormous technical undertaking that I wasn’t equipped to work on.

Plus, this wasn’t even a real problem! The ad dollars were flowing despite the lack of granular analytics and dynamic ad insertions.

My desire to build a tech-driven business steered me away from the path that, in retrospect, seems so obvious.

What I’d do differently

In retrospect, I could have made a bunch of money using a google sheet and an email address.

Here’s how.

I should have started a podcast marketing agency and helped these execs solve their supply constraints and coordination overhead problems.

Here’s exactly how I would do this today. (feel free to steal this idea)

  1. Set up a landing page (using, say, carrd.co), add a “contact us” email address or form
  2. Set up a google sheet with two tabs: Advertisers and Podcasts
  3. Go through the list of execs I talked to and reach out to them. This time, in sales mode, asking for commitments based on the pricing details they shared with me in our initial call. (Advertising is a fairly straightforward industry: if you can get results at a good price, you’re in.)
  4. Once I have a commitment from an advertiser, start reaching out to podcasters until someone is willing to run the ad at a price that makes sense (i.e. profit) for me.
  5. Rinse, repeat, and scale.
  6. (Incrementally automate parts of the process as it scales)


You don’t have to build a hyper-technical product to benefit massively from technology.

Instead, you can use technology to streamline and automate processes within your business and allow you to operate seamlessly at scale.

Don’t get hung up on the tech. Instead, focus on delivering value to end-users as quickly as possible.