Nobody takes a discovery call, everyone takes a podcast invite: how to get customer discovery interviews

Sunny Ray · Feb 17, 2026 · 6 min read

Ask a VP of Engineering for 20 minutes of feedback and you get silence. Ask them onto a show about their craft and you get an hour.

That is the whole trick for getting customer discovery interviews in B2B. Stop asking for help. Offer a spotlight. The interview you wanted happens anyway, it just has a microphone on it.

I watched this play out with a developer-tools startup in February. They had a product working in Python and were pushing into JavaScript and Java. They did not know who their Java buyer was yet. They needed discovery, fast, with exactly the kind of senior engineers who ignore surveys for a living.

Discovery requests ask, invitations give

On our February 10 call, the founder's instinct was the normal one: reach out, ask engineers what they think, learn. Charles proposed something else. Run a podcast for the ICP, and let the pre-interview do the discovery.

I jumped in because I had done four live streams that same day. The question you have to answer with any outreach is what's in it for them. A discovery request puts you first. You need information. You need feedback. You need them to try the product. An invitation flips it.

Just the fact that you want to put a spotlight on them is an easy sell.

That was my line on the call, and I still believe it is the most underrated fact in B2B research. Senior people get asked for things all day. Almost nobody offers them a stage. My estimate on that call was that the chance of getting someone on a call goes up something like ten times when you come with a spotlight instead of an ask. That is my field read, not a study, but it matches every campaign we have run since.

Pick a theme your ICP wants to talk about

The theme matters more than the ask. A show called "tell us about your performance problems" is a survey wearing a costume. Engineers smell it.

The theme we landed on was how AI coding tools are changing software engineering. That is a conversation every staff engineer is already having in Slack. And it walks straight into the startup's pain point: the founder's angle was that AI-generated code can get sloppier and slower, and asking guests whether they were seeing performance problems was the natural way in.

Notice the order. The guest talks about a trend they care about. The trend surfaces the pain. The pain is what you sell to. You never have to pitch to find out whether the problem is real, because they tell you on the record.

On naming, someone on the call had a rule I like: simplicity and gut level win pretty much all the time. Do not overthink the show name. Overthink the theme.

Cutting the list: titles, size, funding, geography

The Java ICP the founder described on February 10 was specific: principal engineers, VPs of Engineering and staff engineers at companies with 200 or more employees, where performance actually costs money. US preferred, not limited to tech.

By February 17 the first podcast campaign was live with a different cut. It targeted CEOs, founders and CTOs at Python and JavaScript companies with 11 or more employees, funded, US-based preferred. The list was about 5,000 contacts and needed cleaning. On that call we added VP Engineering as a title, plus senior data scientists, senior software engineers and data engineers. Leads, not individual contributors.

Two lessons from that list work.

• Start with decision makers, then go one level down. The person who owns infrastructure cost is the one whose pain you want on tape.
• Filter on signals you can see. Funding, headcount and language stack are visible in a database. "Cares about performance" is not, so you find it in the conversation.

One caveat raised on the call: at bigger companies, guests sometimes have to ask their PR team before they can appear. Most companies are fine with it. Build a few days into the booking for that.

The pre-interview is the discovery call

Here is the part most founders miss. The recording is not where you learn the most. The pre-interview is.

In the 15 minutes before an episode you ask what they are working on, where they are getting stuck, what they have tried. You can say "can you elaborate on that" as many times as you want, because you are preparing the episode. It is a discovery interview with better manners.

Then, during or after the show, you can be honest: you described exactly the problem we are building for, can I bring my co-founder onto a second call? As I put it on February 10, it is fulfilling the role of a sales call, which is the crazy part.

Our side was candid about conversion on that call too. Guests from a podcast will not close at the rate of someone who booked a sales call because they were already shopping. The estimate on the call was 10 to 15 percent into clients. But the supply of conversations is close to limitless, and the market tells you things you could not have predicted. The line that stuck with me: after ten years of product-market fit work, you still cannot predict what the market will say. So you put people on a show, ask questions, and find out.

Feed the AI replier real replies first

Once invitations go out, replies start coming back, and it is tempting to switch on an AI to answer them all on day one. Do not.

On the February 17 call the rule was to wait for 5 to 10 real responses before training the auto-replier. You want it to learn from how your actual ICP talks back: the guy who says "sure, when?", the one who asks what the show is, the one who wants to send a colleague instead. Train it on guesses and it will answer guesses.

Also watch the boring stuff. That same week we caught a subject-line bug where a merge field rendered as a literal blank for one recipient. Isolated, fixed, but it is exactly the kind of thing that turns a spotlight into spam.

Where this beats the usual advice

The standard playbook for starting discovery interviews is introductions from your network, meetups, online forums, cold email and cold calls. SKMurphy's list is a good summary and all of it works at small scale.

The problem is that every one of those channels still frames you as the person asking. A podcast invitation is cold email with a gift attached. You get the reach of outbound with the reply rate of something people actually want.

And you keep the content. Every discovery interview becomes an episode, clips, a post the guest shares to their own followers. Their boss sees it. Their peers see it. Your next ten guests come easier because the first ten exist.

If your buyers are senior and technical and allergic to sales calls, stop asking for their time and start offering them a stage. Reach technical buyers with a show.

Frequently asked questions