You've Built a Great Technology, Now What? (My Dilemma)

July 09, 2007

This post is a little different, in that I don't have any clever answer or solution to blog about. The issue is this: I've spent the last few months building a search technology that's a little different than anything I've seen done; it's actually more of a "research engine" in fact. The questions are pretty basic: is it worth something and what do I do with it?

Now I'm not going to lay out everything here in any great detail, but bear with me in my search for answers (a pun). I hope in this process others can learn from my experience.

I've built a functional prototype and wrote a paper on the concept, enough to see it has merit. Most search technologies are focused on finding something you already know something about (buy ipod). What if you really aren't sure what you are looking for? Before you laugh and think this is a dumb question, imagine being a patent examiner looking for prior art, or an FBI agent looking for evidence, or a writer researching a story. In all these cases your knowledge of what constitutes an "answer" isn't well know up front. You need to examine a large body of information with only a vague idea of where to start. You'll know answers when you see them.

That's what my technology is aimed at. I came up with this idea a decade ago but never had the time to work on it. I figured it was likely to be done by someone eventually but so far I haven't seen it. So instead of doing consulting (and making a living!) I took the time to build this to the point of knowing it works and being able to demonstrate it.

The possibilities are fairly obvious but the path is not.

(1) Startup on my own

(2) Startup with investment (VC)

(3) Find a partner (existing search or related company)

(4) Give it away

The major difficulty with all but #3 is that this is a technology not a product. It needs to scale to much larger document collections (my prototype is 18,000 small documents) and needs other infrastructure to make it a viable web application. Once it's scalable it needs tuning to whatever kind of document collections it will be applied to. One thing this won't do is replace search engines like Google; it's much more suited to vertical document collections not generic web pages.

My preference is #3, either as partner or even "captured" as an employee to work on completing it as a product. I contacted Google via their contact form and got no reply, I figured it was a black hole (everyone contacts Google) so it's no surprise. Google would make sense as they already have the worlds largest search infrastructure. There are some 300 search engine companies, not to mention other companies with search needs or products.

I have more than 60,000 contacts via my linkedIn profile which is nice but what are the right contacts? Connections by themselves aren't useful unless you can make a sufficient impression that there is a reason to communicate further.

This is also the issue with #2, I live in the Dallas-Fort Worth area, which as far as I can tell is somewhat of a dead zone for software startups (like a donut hole, all the money and startups are more common on the country's edges). VC's only talk with people via real connections, and of course are mostly interested in startups with a revenue proposition; it's hard to sell something you can't sell to someone. I'm not Ycombinator material and in fact having done a startup before (in the old days) I'm not sure it makes much sense to do it again in this case.

#1 is not a real option, I've already invested what I have just getting to here, and need to make a living again soon. Working after work on something complex is too hard.

#4 is obviously easy and cheap but I'm not there yet; the biggest issue is that I wouldn't get to do much more with it (having a regular job and less time) although it would be nice to see what others could do with it. Even with a prototype it's not really something you just plug in somewhere; the concept is more powerful than the existing code.

Of course I could be imagining value where there is little; however, I have a lot of experience with search, with invention and with just plain thinking different so it's not based on nothing. Once I played with the prototype I could see it was useful; the question was were to go from there and could I make something of it.

Any comments or suggestions are welcome. Email is codistconsulting / gmail.