The Story of Our Failed Startup: The Product

This post is the first part of my story on our failed startup Camargus.

Camargus' technology was born out of a university research lab. We had built a fancy multi-camera system that featured a high quality stitching algorithm to create seamless, panoramic videos in real-time. The image quality was sufficient to be mistaken for a professional camera. People were really impressed by our compelling digital pan&zoom effect. This was possible thanks to an array of cameras which together produced an order of magnitude more pixels than regular HD. We dubbed the pan&zoom effect “virtual camera”, i.e. a software-generated camera which enabled you to look around inside video just like in a 3D video game.

Our virtual camera concept: a panoramic camera array continuously captures a wide-angle view of the entire field, in which you can freely look around using a virtual camera. Either live or in replay mode. 

You could think of so many applications where this kind of tech would be of value, which made it hard to focus and to start developing a real product. This became a lot easier when got noticed by a leading, high-end sports broadcasting service provider called Fletcher. They saw the benefit of panoramic video for sports analysis, in particular in American football. Fletcher ordered a prototype that would be used by a world-renowned broadcaster (ESPN) in paid-for trials during the upcoming NFL season (2010). Our unique selling point: we record the entire field continuously so that you never miss anything. You can replay anything and anywhere using the virtual camera.

It seemed like we were on to something. Fletcher and ESPN wanted to use our prototype to capture revealing shots of contested situations, and even reverse the referee's decision (which actually happened in one of the first trials). But eventually it turned out the frequency of catching something unique was really low. There were just so many regular cameras available already. And there were technical concerns: the video quality was also too far off compared to production cameras and there was an insatiable demand for more digital zoom power. So despite a reasonably successful trial, we were not able to renew the deal for next season.

A setback, but we were still optimistic. We were doing some broadcasting trade shows and our demo would attract leads like a magnet, including leading broadcast executives. Yet we couldn’t quite figure out how to close a new deal. Not everyone is like Fletcher, able to spend $100K’s to build a high risk prototype. But what motived Fletcher exactly? Their dollars did not necessarily represent product value but rather marketing value. More precisely, the trials with ESPN would add to their reputation as a high-tech innovator, regardless of the outcome. It turned out that there are not that many companies like Fletcher.

We had enough confidence to take a leap of faith and invest in a productized version of the ESPN prototype. More time and money went into development. I mean, ESPN had used our shit, right? Fletcher suggested adding more features, and improving image quality and zoom power. They were pulling us to build the ultimate sports analysis tool. But it made us blind for the underlying issue, namely that the product idea was deeply flawed. It did not capture anything of interest compared to the existing camera infrastructure. Could we save ourselves by just making the prototype "better"? And would we sell more than a few units a year then? Probably not. Building a better sports analysis tool (mousetrap?) would have hardly made any waves in the broadcasting industry. It's not disruptive.

There were plenty of ideas to go beyond just sports analysis. One of them was to turn our prototype in a "one-man" broadcasting solution as an alternative to big budget field production. But despite good efforts, we were not able to make it happen. There were technical challenges requiring more R&D, more money. Another hurdle was pricing. We had locked ourselves into the high-end price range. Our bill-of-materials per unit was so high that we couldn't offer a lower price to address a less demanding audience. So we got stuck in a limbo between market segments: not good enough for the high-end, too costly for mid-level or low-end.

So to summarize, we didn't hit a homerun with the first incarnation of our product and we just got lost after that. It's a common issue among startups: product/market fit. I wouldn't go as far to saying that this killed us, but Camargus was also struggling with other problems. I'll get to that in the following posts.