Post-Mortem: openmiq — A crowdsourced podcast platform

This is a follow up from a post a made yesterday: A community to share your failed projects. I thought I would start by posting my own post-mortems to see if there was any interest. If you liked this, comment below or DM me and I'll see about writing a follow up. Enjoy!

openmiq

A crowdsourced podcast builder using audio only voice clips

Type:
Social Media

Tech:
Golang, React, JS, Docker, GCP

Dates:
Feb 2019 — July 2019

What we were building

Hypothesis: Online text-based discussions can become toxic due to the impersonal touch of text. Will adding another dimension (audio) help create more empathetic online conversations?

Using the above hypothesis, a friend and I got together to try and build an asynchronous audio only forum. We enjoyed podcasts at the time and wanted to create a type of democratized podcast. Users would start a conversation with an audio voice clip and then other users would post audio only comments to that clip.

When you clicked on the first audio node, the audio would autoplay each comment ordered by the created date. As you listened, it would sound like an audio conversation between people discussing a topic. At the beginning, we required users to start a thread (like Reddit) with a title and a post, but then we moved to a twitter format where every audio clip was its own "thread".

Early example of what our home screen looked like

How did we get our initial users

When we launched, we tried getting friends and family to post but they were adverse to starting their own conversation (more on that later). Next, we tried our hand at mimicking the Reddit early adoption strategy of seeding content. Unfortunately, "faking" audio was too tough so we tried doing two things: compiling summaries of podcasts and creating our conversation topics.

We posted the podcast summaries to Reddit where we had highly upvoted posts, but people only came to the platform to listen to the summary and then ended up leaving. As for creating our own conversations, these conversations got lower upvotes but more people were willing to participate. It wasn't a lot of people (maybe 3 if were lucky) but it was more than we had before.

How big did it get

We continued with the Reddit strategy and started DMing users on Reddit links to our conversations asking if they wanted to participate. We had some decent success with this strategy where one of our controversial philosophy conversations (anti-natalism) got 20 audio clips in one session! With this distribution channel were able to get to 200+ signed up users by the end (not counting the listeners who didn't sign up).

Why did growth slow?

Even with our moderate success using the Reddit distribution channel, the platform was just not sticky enough to keep people from leaving. Viewing our analytics, we noticed the majority of people who participated in one conversation followed that conversation but didn't join others. Furthermore, we had no new conversations posted by users, it was still just the founding team seeding the initial conversations and posting them on Reddit.

We figured the growth slowed due to these two reasons:

• Posting a public audio clip was not worth the effort. Recording your voice and posting it in public is psychologically hard and takes a lot of effort to produce quality content. Even as the founders we found it difficult to keep posting our own content and would sometimes spend hours on a few minute clip.
• Listening to audio takes too long. You can't scan audio like you would with text which made listening to a full conversation a lengthy process. Furthermore, you can't tell if a voice clip is good unless you listen to it, so it took too long to filter which clips were interesting and which ones were skippable.

Eventually, growth slowed to a point where it was an uphill battle finding users to come to the platform and stay. We had to come to the difficult decision to shut things down.

What did we learn?

This was our first real product we ever built together and we learned a lot from the experience. I had a lot of fun building out the software and there was a lot web audio tech I got to learn. Also, while we weren't able to prove our hypothesis, we enjoyed using it as our guiding vision.

Here's a summary of the things we both felt we took away from this:

• Talk to your users and learn how they use the product. At the beginning, we built things that we thought our users would like. However, when we took time to schedule calls with our users, we learned that they didn't need the stuff we were building.
• Social media products require virality from the start. If people aren't sharing your free social media platform, then you're not likely to succeed. With our product, there was no word of mouth, and with no word of mouth, we just couldn't grow.
• We liked hearing our friends voice more than a stranger's voice. During the times our friends would post a clip, we tended to disproportionatelyenjoy their content over a strangers, even if it was objectively a "lower quality" post. For us, we found it was more fun creating content to share with our friends.

Overall it was a fun experience and we were able to learn some key features of product building. With these learnings, we ended up pivoting to another project that was still within the audio space. But that will be for another post ;)