For once, a post with the word Bloomberg in the title has nothing to do with New York’s (thankfully) outgoing Mayor. Recently Bloomberg L.P. got in trouble. Bloomberg reporters were relying on searches performed by customers in order to report on developing news stories. At the time, I noted that when a data provider starts to compete as a service provider, access to huge droves of big data becomes problematic.
Last weekend, I was talking with a technologist at a legal conference about how big data could transform law. I posed a hypothetical. Wouldn’t it be extremely informative for a company to parse through the searches a lawyer makes to try to figure out what he is doing, how he is doing it, and what the strategy may be? This is how Google and other companies builds profiles of behavior. The technologist said that a company that relies on the trust of its users would never do something like this. I raised the example of Bloomberg. Recent revelations suggest that this behavior was not aberrant:
But interviews with more than 30 current and former Bloomberg employees paint a picture of an aggressive, hyperkinetic organization that not only tolerated but encouraged an unusual symbiotic relationship between the company’s news operation and its business interests, including the use of the terminals to break news.
Many of these reporters say they routinely used the terminal’s function, called the UUID command, to find background information on subscribers, including contact information, when the subscriber had last logged on, and weekly statistics on how often customers used a particular function, like equity shares or currency markets.
Reporters also say they talked to their business-side colleagues to gain early access to company announcements, including company releases and bond prospectuses about to be made public, and queried Bloomberg’s own help desk to find out what pages and features customers might have been using.
“We were told again and again and again, find ways to use what’s on the terminal to write stories,” said one former Bloomberg reporter who, like most employees, was required to sign a strict nondisclosure agreement that prohibits them from speaking about the company. “They never said ‘Oh, please be careful and don’t breach any kind of privacy,’ ” the former reporter said.
A senior Bloomberg executive not authorized to speak for attribution confirmed the reporter’s account. “It’s not like this was some deep Bloomberg secret,” this executive said. “If it’s on the terminal, we’d say go for it.”
The temptation to look through this data is so great, that a company with such an institutional reputation as Bloomberg gave in. What about other players in the legal field? One item that I am exploring is how a law firm that relies on analysis from big data could comply with the rules of ethics–specifically conflicts of interests, and relying on disclosed data. More later.
Update: More on how only 13% of the homepage on a Google search is actually search results (influenced, of course, by Google’s algorithms). 29% is Adwords, 14% is navigation bar, and 7% is Google Maps.
Google.com’s search results have all just become links to Google’s own services.
Google will always be indispensable as a service, no one doubts that, and it’ll always feel the quickest but it’s become a wee bit more complicated and self-serving than old Google. Google.com is just a way to round us up to use more Google.