Startup lessons learned - start with a problem
Start with a real, concrete problem to avoid needless headaches on your founder journey
When we started EnVsion, my co-founder and I had an avid interest in how AI could be applied to video in order to extract some type of insights. Given my background in mobility and fascination with AI and autonomous vehicles (AVs), I was drawn towards the growing popularity of dash cams because I thought we could build the equivalent of YouTube for dash cams, with the added AI analysis component to identify and explain accidents in those videos. I wasn’t sure what the end goal was nor what the business model would look like(e.g. would we sell our data or insights to insurance companies or AV companies?).
We hadn’t identified a real problem at the time. From this bad start it took us much longer to correct our course and get onto the current road we are on.
Interestingly, when we started the company we had brainstormed a few ideas that had video at the core. One such idea was to auto-transcribe videos and make them searchable. You see, at we were working together as consultants for a big energy company, building a new type of predictive analytics platform. We spent countless hours in video calls speaking with customers to understand their problems. We often wanted to go back to these recordings to extract key insights from these conversations to share with the team or stakeholders, and ultimately make more informed product decisions.
Doing this was too time-consuming though. It was also soul crushing. Basically, we had to:
remember the name of the video file of interest
watch the full video to identify the start and end of key soundbites
write down these timestamps and use a video editing tool to create clips at those specific timestamps
It seems we had a real problem here, but we just felt the AI to explain accidents route was cooler and conducive to building a multi-billion dollar business. I certainly thought I was smart enough to figure the shit out of the uncertain and unproven idea we initially went for. It sounded grand… and hollow.
Hindsight is 20/20, and I get 21 marks for saying we should have stuck with the problem we were familiar with. That’s eventually what we did. Due to COVID-19, we also witnessed the democratisation of remote-first work models in the business world. These changes in the way people work is in great part powered by video. We began swimming in this favourable tide: we spoke with many founders and employees in remote companies to understand the scale of the opportunity. We then realised that the initial problems from our consulting days fit neatly into our newly formulated vision to unlock the knowledge in companies videos. I expand on this vision in another blog post here.
So now we are building towards this vision. Our starting point is unsurprisingly UX Research, because that’s also where we were struggling back in our consulting days. There is still much to learn for us in this space, but we are becoming increasingly familiar with the domain. Finally, it seems we found our problem.
Reflecting on these self-inflicted wounds, I say to my younger, greener, founder-self: “It‘s obvious, you should start with a problem”.
It certainly is now.
Thanks for reading this post! I share my learnings and progress while building my startup EnVsion! At EnVsion, we’re creating the central repository for UX researchers and product teams to unlock the insights from their user interview videos. And of course we use AI for this ;).
EnVsion is for you if you’re a UX researcher or product manager who is overwhelmed by the hours of video recordings from calls with users and customers. We help you highlight the key moments from these calls so that you and your team can make better product decisions.