Tuesday, December 4, 2007
Intellectual property in a business like ours typically refers to software technology. Code, algorithms, patents, methodologies – all required to recreate and build the product and protect it under IP laws.
In the business of search-driven research, it turns out that models are as significant a part of our IP as the traditional software code is. We have an excellent search engine and surrounding systems, but more importantly we have models of what’s interesting to the professional user. For every “topic” that’s to be detected we build adaptive market models which characterize the company or topic the user is interested in seeing - and because we have their portfolio or relevant topics we show high precision in selecting only documents that are of interest to them. (These we then send in a daily report of links and summaries or store in the online user database.)
We did not have this all figured out in the beginning. FirstRain is a classic case of working with a market to determine where the value is and then continuously course correcting the product until users say “yes!”. When we made the strategy switch in early 2006 to focus on investors first and then come back out to C-level executives we had a hunch that the topics (i.e. the models) that we would develop would be interesting but we had no idea how big a part of the secret sauce they would become.
Today we have team of analysts – over 100 people and growing – who develop and manage the analytics IP. We have developed high precision models for
- companies – representing 95% of the world wide trading volume
- industry topics
… and the list of types keeps growing too. The more we work with the buy-side and our corporate clients, the more we see the richness of experience we can produce if we can only model what’s critical and filter out the rest.
And, no question, industry topics produce the most fascinating results. Knowing what investors care about on a security, when the ticker or company is never mentioned, is what produces the cream of the information.
The biggest lesson for me and my team was that we need to always pay attention to where the IP really is. We had done a great deal of work on analytics but had not catalogued it sufficiently well when I had a eureka! moment - realizing that the analytics IP was as fundamental as it is. That’s fixed now and a great side effect of the realization and subsequent work has been the excitement and ownership that’s now developed in the analytics team.