In previous Industry Insights, I've used specific movies ("Jerry Maguire", "Moneyball") to explore key themes in business intelligence (BI). Let me continue that by using the 1992 hit movie, "A Few Good Men".
Apart from being revealing about my age, there is an iconic scene in the movie between Jessep (Jack Nicholson) testifying and Kaffee (Tom Cruise) cross-examining, which illustrates how BI should challenge conventional wisdom and find previously undiscovered 'truths' as opposed to just delivering reports.
Jessep: You want answers?
Kaffee: I think I'm entitled to them.
Jessep: You want answers?
Kaffee: I want the truth!
Jessep: You can't handle the truth!
Dramatic? Yes. Memorable? Absolutely.
It was brought to mind by a recent analytics conference I attended, where a plainly frustrated audience member asked: "But how do we get executives to realise the benefit of BI?" I suggested that in order to get the attention of BI customers, instead of providing the packaged reports (while essential), which form the basis of most BI systems, the data needs to be cross-examined to find the unknown truth that challenges mindsets and realises measurable value in the business context.
Let's take a look at an example.
Credit crunch
In the years leading up to the 2008 credit crunch, there were a number of facts which were consistent and known. US mortgage credit was being extended at a rate not seen for decades; house prices were increasing at a faster rate than ever before, outpacing earnings. Sub-prime (or higher risk) mortgages were becoming an increasing share of the market, interest rates were at multi-year lows.
This was all data freely available at the time to interested parties. What was interesting was the different "business intelligence" approaches used to derive conclusions from this. On the one hand, there were banks such as Lehman and Bear Sterns (among others), which made huge profits leading up to 2008 by betting that the housing market would not collapse based on prior history. In addition, significant executive bonuses were paid on the basis of total sales (and not the quality thereof).
Depending on point of view, people may come to different BI conclusions from the same data.
On the other hand, there was the hedge fund manager, John Paulson (among others), of Paulson and Co, who, when looking at the same data, came to the conclusion that a housing bubble had developed which was unsustainable. The fact that the mortgages were rated as safe by rating agencies meant the insurance against default was being sold at low rates (a simple analogy would be being able to buy insurance on a Ferrari that someone else owns for R1/month - it will cost R1/month; but, it only takes one crash to be able to collect on the value of the Ferrari). Buying up large quantities of this insurance, and short-selling mortgages, earned Paulson's hedge fund $15 billion in 2007 alone. His own earnings that year were $3.7 billion.
It would be fair to say that Lehman (bankrupt) and Bear Sterns (sold) were losers in this BI game, and Paulson was a very big winner. It's a short step from here to making the case to executives that real BI matters.
E-tolls
Closer to home, it's fair to say the proposed e-tolling system is top of mind for residents, and that depending on point of view, people may come to different BI conclusions from the same data. (It's beyond the scope and courage of this author to pronounce on the rights or wrongs of e-tolling itself, merely to consider a data analysis case).
Sanral suggests that based on its data, 82.8% of road users will pay less than R100 a month.
Using the gantry rates published in the Government Gazette, and assuming a tagged standard passenger vehicle (class A2) commuting to work 21working days a month, with an average gantry costing R2.77 for a single return trip, means every gantry passed will cost R58.17 a month, excluding evening and weekend trips. This means the average vehicle commuter only needs to pass two gantries on the way to work to exceed the R100/month.
The interesting thing about this is that the definition of 'road user' may allow for stats which seem to indicate a low average cost. If included in the term 'road user' is all passengers, possibly cyclists, and motorists who do not use the highways themselves, then it might be feasible that 82.8% will not pay R100, but the average motorist seems unlikely to avoid less than R100 in monthly charges.
The point here is that the story from the data is often just a point of view and not necessarily an absolute truth.
When it comes to considering the value of BI, the tools and the reports are less important than the cross-examination which reveals an often unexpected and valuable truth about the way people view and perform their daily business tasks. If customers want answers and can handle the truth, companies owe it to them.
Share