One of the perpetual human foibles is to think that your world is a microcosm of the larger world - and to be surprised when it isn't. The Americans wish the French weren't so French, and the French wish the Americans weren't so American. You get the idea.

This principle applies to professions as well. IT workers - successful in an environment that rewards logical thinking and attention to detail - are sometimes amazed at how much sway politics and gut feelings have in the user world.

This is especially true in BI, where the data is structured and categorized. In the world of OLAP cubes and data warehouses, it's easy to think that your work is done when you deliver a repository full of accurate and relevant information that the business can use. And, in the strictest sense, it is.

However, in this column I'm going to argue that it isn't. Recent studies have shown that humans are a lot more emotional in their decision-making than we'd previously thought - or care to admit. Consequently, any IT professional involved in helping the business make decisions needs to take these findings into account. By reviewing what we now know about human decision-making, you can make your BI initiatives more useful.

Deciding in a Glance

Blink: The Power of Thinking Without Thinking, by Malcolm Gladwell, is a well-known book which popularized the view that decision-making is complicated. In it, Gladwell argues that in many cases the best decisions are the gut reactions - the mind is somehow able to synthesize what it has learned over the years and just "knows" when something is right or wrong.

His first story is about a kouros, a Greek sculpture that the J. Paul Getty Museum in California purchased. All the "hard" facts seemed right: pages of letters documented who had owned it previously, and a geologist from the University of California said that it was made from marble from the island of Thasos and that it was old. However, experts in Greek sculpture thought something was wrong the instant they looked at it. They couldn't quite pin it down, but it didn't look like similar sculptures they knew. Ultimately, the statue turned out to be a fake. Someone had forged the supporting paperwork and artificially aged the marble.

Looking at Too Much Data

Another example that Gladwell mentions is predicting which patients experiencing chest pain in the emergency room at the Cook County Hospital in Illinois would actually have a heart attack. Experienced physicians were doing their best, but statistically, the results were all over the map. Brendan Reilly, chairman of the hospital's department of medicine, decided to compare the physician's performance with that of a decision tree, looking at only three risk factors. At the end of the two year experiment, the results were conclusive: the decision tree was 70 percent more accurate than the doctors. It turns out that the physicians, in their quest to get all the facts, looked at too many variables and got confused by the "noise"; the formula, on the other hand, looked only at the most important predictors.

Decision-Making Isn't That Rational

Gerald Zaltman, in his book How Customers Think: Essential Insights into the Mind of the Market, notes that customers' decision-making processes are not always logical and conscious. He talks about several findings: thought is based on images, not words; metaphor is central to thought; and most thought, emotion and learning occur without awareness.

Zaltman also points out that "there is nothing more subjective than a number. The only objective aspect of a number is its derivation (as through particular calculations). Before ever 'crunching numbers,' managers make crucial judgments about which target markets to sample, how to sample them, when to do so, what topics or questions to explore, and which analyses to perform on the data. All of these human judgments affect the actual results obtained."

Don't Throw Up Your Hands - Just Question a Bit More

All of these findings and stories are unsettling to people who believe in rational decision-making. So what's a BI professional to do - just throw up her hands and move into another profession? Not at all, but it does make sense to double check with your users about how much value you're providing.

For example, if you've rolled out a dashboard to help busy executives make decisions, sit with a couple of them (separately) and ask questions such as:

  • We put up this dashboard nine months ago; are you comfortable that those are the best metrics? Are there any that you now realize aren't very useful or are there new metrics you'd like to see?
  • Other than the dashboard, what information do you look at before you make your decision? (You may find that rather than serving as the one-stop-shop for better decision-making that you pitched, it's really one of six spots users look. Investigating the other sources that your clients use may help you to figure out ways to make their quest easier.)

Supporting stories are another option: rather than just depending on a screen full of sales numbers to describe and predict sales success, maybe augmenting them with a group of 30 "How I made the sale" written stories from sales reps will help sales managers get a better feel about how "firm" those numbers are.
Remember, your ultimate duty is to help your company make better decisions. In the end, going beyond the numbers may be what's required.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access