(DRAFT AS OF THURS APRIL 26, 2007)
Transparent Bundles
I started this blog three years ago, in the spring of 2004. I was working on a new research model for hedge funds that told stories using data, culled from the Internet. We did not believe that the god-analyst structure of the pre-bubble tech brokers (Salomon/Grubman, MorganStanley/Meeker, MerrillLynch/Blodget…) led to compelling arguments and rigorous convictions any longer. It seemed like the $3-5 billion stock trading commision market wanted to support a new kind of research that was based on massive amounts of observed data rather than a single person’s prediction.
Or at least this was what I learned by going door to door to meet with hedge fund analysts. I was selling an algorithm for turning a certain kind of input (ie online behavior) into an equally certain output (ie real-time analyses of the consumer equity sector). The likely customers, Tony and I assumed, were Portfolio Managers at small funds that had concentrated positions in companies like Best Buy and WalMart, not to mention Expedia and Amazon. Most of these investors were working in mid town offices between 3rd and 5th Avenues, above 42nd and below 59th street. The more I started selling to this new generation of Wall Street, the more I realized how little time I spent on Wall Street per se.
In the Fall of 2002, while Spitzer was attacking the sell-side and Regulation Fair Disclosure was taking hold, I began telling the story that to understand the behavior of the American consumer, you needed to understand what she was doing on her computer:
It only took a few minutes and I had the analyst nodding in approval. They bought into the methodology, but wanted to know what the deliverable would be. Our initial input was raw comScore data that we had licensed exclusively for sale to the buy-side. This substantiated the “proprietary data” story and enabled us to offer a regular report called “Consumer Macro” that drew from this rich panel of live user data to produce real-time insights into the mind of the retail economy:
We were successful in getting investors to try out our new service and we started to learn how to take data that had been mined from the Internet and refine it for inputs into their Excel spreadsheet models. We also expanded beyond simply licensing 3rd party proprietary data to begin capturing data directly from the Web through bots, spiders and what have you. In disintermediating the traditional research providers, we were intermediating ourselves between the online individual as data producer and the buy-side fund who was buying statistically significant, aggregated insights.
In the Spring of 2004, we were fortunate to get featured on the front page of the Wall Street Journal as representative of a new kind of Wall Street research:
Increasingly, Stock Research Serves the Pros, Not ‘Little Guy’
May 2004, by Ann Davis Staff Reporter of The Wall Street Journal
In early September, CarMax Inc. told investors it had met used-car sales targets for a just-ended quarter. But a small number of investment pros would soon learn there might be a catch.A stock-analysis boutique, Majestic Research LLC, had dissected car-registration data that it regularly gets an inside look at. It discovered that CarMax’s sales for the second quarter apparently hadn’t grown evenly across all dealerships, but had spiked at a single California outlet.The research firm sent out a flash alert. Some big investment funds quickly sold or changed their bets on the shares. Several weeks later, CarMax lowered …
