Would You Like Fries with that? Reflecting on the Value of Cheap Information

Larry Neal’s recent column in September’s Decision Analysis Today talks about the value proposition of Decision Quality (DQ), and a key point is the Value of Information (VOI). One of the most powerful things about decision analytic methods is the ability to explicitly calculate the value of information, in order to make an intelligent choice about whether to buy it. Everybody loves information, and as Larry points out, the pre-DA/DQ inclination is often to sweep handfuls of it into the shopping cart, as long as none of them look expensive. While price matters, with information as well as everything else, so does value. It’s remarkably easy to talk yourself into buying cheap but worthless information.

Cheap Information

Examples of worthless information in business abound. An automotive client of a firm I once worked for used to habitually perform customer acceptance tests – basically, put real people in a car you’re going to introduce next year and see how they like it. Isn’t that a good thing to know? Well, yes, if you know it early enough to change the design of the car. But designing a modern automobile is a massive multi-year engineering effort, to say nothing of retooling the plant. A few months before introduction you’re not going to change anything of consequence – maybe you can tweak the option packages, but even a proverbial restyling of the taillights would be pushing it. So why do the test? Well, it had to help that it was cheap – once you have prototype cars, just drag a focus group out to the parking lot.

There is a special kind of nonlinear price sensitivity that comes into play with information. Cheap is good, but even better is free. How often is information free? Well, in business, basically never. Sometimes the information looks free, the result of desktop research or querying suppliers. However, time is money, whether it’s analyst work time, or simply the time value of delaying the project. For a high stakes investment decision, a one month delay might have a cost in the millions of dollars, yet managers are never as parsimonious with time as they are with cash. Well managed pharma R&D projects often try to get a handle on this by calculating the cost of a one month delay. Even if it’s a very rough estimate, that number can help inform decisions on the ground. Are we really going to delay this project two months because we’re understaffed in regulatory?

Organizations that are gluttons for “free” information can often be recognized by their agonizingly slow rate of new product introductions. While their competitors are taking a few shots on goal, they’re still studying the defense. This is not just an issue in business, but also in the public sphere. Sometimes the best way to study something is to do it – go out and paint some lines in the street.

Fortunately, as people who know decision analysis, we have the ability to do the math. If you have a model, you can easily calculate the VOI and see if it clears the (possibly low) hurdle. But even if you don’t have a model, such as when you’re making one of those countless small decisions that don’t warrant much analysis but benefit from DQ thinking, you can always do an armchair VOI calculation. Is there any conceivable way this el cheapo information would change my decision? If not, it’s not worth $2.