Raw Audio Experiments: Benefits and Drawbacks

by Ben Fisher

Podcasts offer some of the best content out there.

What better way to capture wisdom than to round up smart people, then get them to talk about what they know?

The same can be true of one-on-one conversations between experts. The problem is: no one’s recording, and no one’s listening.

My friend Stuart and I decided to change that. An experienced podcast host and founder of Procket, Stuart called me up and we had an unscripted conversation about raw audio. Then, we decided to go meta and take our own raw audio and share it with the world. Here’s what we talked about.

Podcast recordings and “lost” raw audio

I recently started a new podcast with Adii Pienaar called the Dog + Pony Show - a pun based on my business 1-Click Pony.

Adii and I wanted our podcast to be unique. Instead of taking a “behind the scenes” approach of building a SaaS business, we planned to focus on building content for our target customers.

With a similar customer base (Shopify stores) and at roughly the same stage in our business journey, we soon realized that we have a lot to talk about.

We recorded our first 25-minute episode - but we also spoke for an additional 60 minutes before and after the formal Q&A.

I was bummed that we couldn’t use all that great audio. It wasn’t structured, and we mentioned a few confidential details like client names. But the fact remained that I was just as impressed and inspired by our pre-show chat as the published version itself.

Hidden themes in raw customer audio

When I told Stuart about my new podcast venture, it immediately resonated with him.

In his work with customer interviews, Stuart often sets up private podcast feeds so that companies can listen to customer interviews firsthand.

He used to think it was enough to just send a report with key takeaways and insights. But in reality, often different listeners clued into different themes.

“So many times now different people on the team will have totally different reactions to the exact same audio,” he told me. “It's better that you have a diversity of reactions to the same input.”

As we talked about the value of raw audio, we realized that there were other interesting applications and arguments beyond just podcast recordings.

Recording meetings: yay or nay?

I must have an obsession with recording, because I also record every customer meeting I have.

After I record a meeting, I send it off to my team - raw Zoom footage, full transcript, and all. Then, they can listen to it at 2x speed.

As an engineer, I thought it was important for them to hear the customer firsthand.

The con? I just read the book The Mom Test. Apart from lots of other great advice, the book said not to make your team listen to every meeting. Because ultimately, they really won’t listen to the whole thing. Instead, you should summarize the major points for them.

It kind of left me with a catch-22. I don’t want to be the business leader where everything’s in my head (a classic bottleneck), but I also know that people don’t have time to listen to every meeting, even at 2x speed.

Raw audio helps you delegate with questions, not decisions

I think there’s a happy medium to be found - one where I provide a high-level summary, but leave the team to interpret exactly how to move forward and flex their solution-generating muscles.

I don’t want to be a dictator-style leader where I lead with decisions. I do want to be a democratic-style leader where I remain receptive to different points of view.

Having access to the original customer audio will ensure that my team is empowered to understand the context.

There are so many examples of companies where the CEO is the decision-maker and they drive their teams crazy with shifting priorities. One day it’s one thing, the next day it’s something else.

With only so many hours in a day, this is something I’m still figuring out. But no matter what I decide on, my team will still have access to raw audio as part of our routine.

Consensus is the enemy

Just because you share everything with your team and invite them into the process doesn’t mean that you need to aim for 100% consensus in everything you do in your business.

In fact, Stuart brought up the point that consensus can be harmful. “The last thing you want to do with a product is built around averages,” Stuart told me. “Getting to consensus is not always helpful.”

This echoes the opinion of David Cancel, CEO at Drift, who said that building around averages (consensus/agreement) results in an average product.

It all comes down to the sweet spot of trusting your team, giving them access to all of the raw information you have, and then having the balls to make the final calls.

Does your team use raw audio in any capacity? I’d love to hear how.

Postscript

Tool-wise, Descript's the winner so far. It's best at detecting multiple speakers + accurately transcribing what's said. The tools are pretty useless without that. Unfortunately it's also a laborious workflow, so I'm still searching for a better solution.

I just got into Vowel's beta — I haven't tried it yet but am cautiously optimistic.

Here's what I've tried:

  • Meet in Zoom => transcribe via Descript => put into Notion doc w/ takeaways => share w/ team. GOOD BUT LABORIOUS.
  • Meet in Zoom => transcribe via Quen.io => put in a Notion doc w/ takeaways => share w/ team. FAIL -- inaccurate transcription + No way to relabel speakers.
  • Meet in Zoom => auto-transcribed via Fireflies.ai => auto-posted to #meeting channel in Slack. FAIL -- inaccurate transcription + No way to relabel speakers. Soundbites feature is interesting + its AI filters sound.
  • Meet in Vowel => auto-transcribed w/ any notes => auto-posts to #meeting channel in Slack. TBD.