Who are the real Moneyball heroes?
The movie “Moneyball” has been pretty endlessly worked over by bloggers who are attracted to the story of how data triumphs over expertise. Heck, I even wrote one myself when I had the chance to meet Billy Beane at a conference.
So this won’t be just another excuse to post an image of Brad Pitt and make the observation that you can make better business decisions by running the correct algorithms. This isn’t about Big Data, faster analytics or mathematical models: it’s about the data gatherers themselves.
In the baseball story, the data gatherers are the nameless guys hunched over laptops, typing in each detail of every play, for game after game. Where was the ball pitched, was it a grounder or a fly, how was it fielded, etc. Amazingly it was done to this level of detail long before laptops. In fact, the data was gathered a long time before it was used in a structured way. Englishman Henry Chadwick –who was raised on cricket– developed the box score in the 1870s (and chose the letter “K” for a strike-out).
Without accurate data, there is no mining
Only once the data gatherers have done their work is there any data for the analysts and modelers to play with. The data-miners may have degrees in mathematical modeling, but Big Data can only take off once there is Data to actually work on.
The Data needs to be accurate, and normalised. In baseball it is pretty easy to identify a player by name, e.g. “Derek Jeter”. Jeter doesn’t ever change his name, but companies change their names all the time such as Rim becoming Blackberry. An analyst trying to answer “how much do we spend with Blackberry a year” must hunt through records for “RIM”, “R.I.M”, “Research in Motion” and many others.
Business networks are the data gatherers, and we are starting to develop the box scores of internet commerce. Instead of pitcher and batter, we know statistics of who is buying from whom, at what volume and transaction count. Just as in the Baseball Box Score, or the cricket Score book, each tiny interaction needs to be recorded so the larger picture can emerge. Business Networks are also becoming the holders of “canonical” data on companies, tracking name changes, mergers, Joint Ventures, cross-holdings and subsidiaries.
For example, say you are responsible for the Purchasing Operations at a bank, and your spend analysis tells you that you have 437 suppliers of Facilities Services. A business network can tell you how many suppliers the average bank has, allowing you to benchmark your supplier rationalisation program.
Or as a seller in a new market, you will be concerned about cash flow, especially how quickly your invoices are likely to get paid. The Network could tell you what the median DSO (a measure of invoice payment speed) is in this market.
Business networks are carefully filing away those box scores waiting for the statistician who will start to develop the algorithms.
And maybe Brad Pitt will play you in the movie.