My friends over at Juice Analytics had a post a little while ago that discussed customer analysis in the context of a recent article by Malcolm Gladwell (the author of Blink) in the New Yorker comparing puzzles (where more data is the key to finding a solution) and mysteries (where the data is available but drawing the right conclusions from it is the challenge). The Juice folks say that customer analysis is a mystery (needing understanding and analysis) and not a puzzle (that needs data) and I have to say I agree with them. Most organizations have lots of data, what they lack is insight or, if as Dave McComb says "Information is data that changes you" then one could say they lack information. Add to this the fact that many of the folks trying to solve the mystery of how to treat a customer are under time pressure and in high turnover positions, and the need to turn the data you have into actionable insight - to solve the mystery of how to treat this customer - becomes even clearer.
The second interesting consideration here, suggested to me by a colleague (thanks Dave) is whether this article argues against automated modeling driving decisions.There is a fair amount of interest in automated model building and adaptive models (like Siebel Real Time Decisions). While there are various pro and con arguments that can be made, the article prompted this one: Without understanding the context of the problem, and the context of the data as it changes, the likelihood of arriving at spurious results increases.
The young journalists in the Watergate story could solve a puzzle but not, perhaps, a mystery. Automated analytics can solve puzzles, but customer treatment is a mystery and for that you still need someone with deep expertise.