WalMart, like the great Presidential Divider not Uniter, inspires great praise and ire at the same time. WalMart for bringing lower prices to consumers plus shopping and job opportunities to disadvantaged areas but at the same time creating deserted downtown cores and helping export domestic jobs to lower-labor cost foreign suppliers and economies. But there is one arena where WalMart gets consistently high marks – IT Systems. And one IT arena WalMart has had a lasting impression is in the BI field.
For close to 20 years, Wal-Mart has been at the state of the art and/or defining the IT standards in real time data collection: POS-Point of Sale instant data capture, inventory monitoring and control systems, shipment control and tracking programs, handheld devices for inventory monitoring, plus lately the planned use of RFID-Radio Frequency Identification Devices for tracking and controling shipments of goods throughout WalMart. Part of this has been devoted to improving daily operations in each area of the business. But another goal has been to constantly improve the BI-Business Intelligence that WalMart buyers and store managers have at their disposal for the choice and pricing of goods on sale at WalMart.
WalMart advertises always lower prices; but the real name of the game at WalMart is identifying consumer buying trends better than the competition and then having a buying and pricing strategy for every situation to take advantage of those ever changing trends. WalMart is in the business of being the smartest buyer in the region, nation, world – so it can be the most profitable seller at its avowed ever lower prices. And better BI is squarely behind this success.
But WalMart also holds its BI cards pretty close to the vest. Example, at a VLDB-Very Large Database conference a few years ago, I was seated next to two WalMart IT types and they were taking notes and commenting on their experiences with the conference speakers ideas. As speakers touted hundreds of gigabyte databases(“been there done that – whose trying clever clustering now”) or huge dimensional databases (“yikes, watch out for dead end city”) or realtime reporting (“now you are talkin -but what about Web approaches”) – there was a running audible commentary. Until I made the mistake of saying that I could not help overhearing .. and what did they think of new ROLAP engines?
Cordial reply. And further queries brought generalized responses – “well we havent really tried dimensional databases so we really cant tell yet. But what do you think of the dimensional approach ?” And as for overhearing useful impressions – it was now down to knowing glances and curt comments – “thats interesting” or “wonder what the Info guys would say”. Another example of the Heissenberg Uncertainty Principle in the Large – in the process of observing a system you may add even the tiniest energy but that can change the dynamics dramatically.
Now WalMart IT is not ultra-secretive – after all they have to interface with hundreds of suppliers and support agents setting various cross platform APIs and ways of exchanging data. But neither are they close to promiscuous about what they are doing. So it was interesting when Information Week featured a cover story about WalMarts CIO weighing in on the outsourcing question. Now this may be window dressing in part because WalMart is under press scrutiny for favoring low cost overseas manufacturers and suppliers; but WalMarts CIO blessed the notion of not only building their own IT systems but using domestic talent primarily for the need to continually finetune such systems but also to prevent their innovations from being made quickly available to competitors.
So the old Wall Street maxim applies here – When WalMart speaks about its BI practices – do take notes.