I'm somewhat reluctant to post this as it is by far my largest personal failure. However I see a lot of comments on HN about the negative value of business guys and I would like to post a counter example from my own experience.
The initial technical idea was actually mine - with this new-fangled 802.11b stuff now available on PDAs they would be perfect for VoIP. Any bandwidth problems (in a dense area) could be solved by using 802.11a back haul to a network of fixed location seed nodes, then all the handsets could be part of an 802.11b mesh network.
Obviously I had no way to implement this. I was a first year undergraduate and you can't build much of a mesh network with one PDA you bought by raiding your food budget. One day I got talking to a guy on a train about this who got pretty excited about it. A week or two later he called and asked if I would mind putting together a paper about how it could work.
Some months passed and every now and again we had a short conversation about one thing or another, but nothing much seemed to happen.
One day I'm asked if I want to get going with building it. Based on nothing more than a short technical paper (written by an undergrad!) and a few power point presentations my contact had successfully acquired sufficient funding for expenses and prototype equipment and obtained a lead customer who was willing to pay enough to get us up and running. The customer had also loaned us a large site to use as a test bed. There were also a few follow on leads at a reasonably advanced stage.
I dragooned a friend into working with me and he did most of the work getting the prototype up and running. However in practice with real world interference (hello, steel framed buildings) 100 meters was about the best separation at which voice calls would still work. By very careful placement of the static nodes we managed to get pretty good coverage on our demo route but it was never going to scale to a large operation. My great idea was not viable because my assumptions were faulty and I had significantly oversold the capability of a dense mesh network to overcome that.
At this point things started falling apart slightly with our lead customer. Our business guy did a good job at keeping them happy until it was obvious that things weren't going to work, at which point he and the company switched tack to something else.
In this case our business guy provided the opportunity and the push to take a shot at it - in an odd reversal from a lot of start ups it was only the technology that could let us down.
The things I regret most from the experience are:
- Being a pretty bad technical lead. I'm pretty sure my friend wouldn't want to work with me again, which is the worst kind of bridge to torch. I have put a lot of effort into acquiring the right skills and approach to make sure that doesn't happen again.
- Not having validated the technical side more thoroughly. If I had known it wasn't going to work I could have had that kind of push behind something with a chance of success.
- Not having put enough effort into maintaining contact with our business guy. He was pretty good at what he did but I don't think I appreciated that at the time.
Now I'm doing my own thing (details in my profile - half way through month 2 now) and most of the things I have to learn are on the marketing and sales side. I'm really enjoying it but I totally underestimated how much work was involved in doing it properly.
Happens all the time. At my last startup we added a new product line based on work by this super-smart PhD who had been in the business since he worked at ThinkingMachines. His product worked great -- for one really narrow use case for a single customer. His claims to the contrary, it ended up being so brittle, so poorly coded, and so ungeneralizable we had to pull the whole product. Luckily it wasn't material to our eventual acquisition :)