Features

Q3. GPS: After you left Microsoft, you founded digiMine (in 2000), which provided Data Warehousing, analysis and data mining applications via the Web. What was your biggest success there? Biggest failure ?

Usama Fayyad: We started DigiMine in March 2000. The idea was to evolve a hosted data warehousing and data mining on-demand service since most companies were unable to build their own systems effectively. Our biggest success is our unexpected traction with large Fortune 1000 companies. We quickly built up a big business with some of the world's largest companies in several verticals including Financial Services (e.g. American Express, GE), Manufacturers (e.g. DaimlerChrysler, Ford), Telcos (e.g. AT&T, T-Mobile), Technology Companies (e.g. Microsoft, Palm), Publishers (e.g. Wall Street Journal, CBS) and Retailers (e.g. Barnes & Noble, Nordstrom). The speed with which that distinguished list of clients grew was incredible and is something I am very proud of.

Looking back, we probably underestimated the costs and complexities of the business. While the business is interesting, the fact that Business Intelligence and Data Mining are not well-defined and well-understood by the market, led us into situations where expectations of clients exceeded what technology could deliver. So even though we were deploying solutions that were more successful than what clients tried to build themselves, the lack of maturity in this business required us to build a growing Professional Services (Consulting) arm. While that business is very profitable, it is not scalable.

Ironically, this led to the ultimate creation of DMX Group which focused solely on the consulting business. The consulting also led to a very carefully chosen refocus of digiMine into one of the verticals we were operating in: ad targeting for online publishers and the evolution to Revenue Science today -- which is a strong player in their field. In the transition, I transitioned from being President & CEO of digiMine for over 3 years to becoming Chairman of Revenue Science and focused my energy on growing DMX Group.

DMX Group was thus a "spin-off" that absorbed the breadth of the mission of digiMine but was not pursuing the hosted business aspects. It was a wonderful situation because I had a book of business from day one, a profitable practice, and a company that was entirely owned by me (no venture capital needs). The allure of DMX Group was the pursuit of the toughest technical data mining challenges but in a context where the business users and executives had to clearly understand and value to work. In that sense, it was a unique effort.