Data Uncertainty Principle

by Don Harkey

I'm going to "geek out" on you for a minute, but bare with me because I think it will be worth it!

In advanced physics, there is a principle known as the Heisenberg Uncertainty Principle. The principle states (oversimplified) that you cannot perfectly know both a particle's position and velocity. The problem is that by simply observing the particle, you are impacting either its position or its velocity.

In management, I would like to submit another theory called the Data Uncertainty Principle. This principle states that any data destined to be used for a predetermined purpose will become unreliable.

In other words, say you decide to use your inventory dollars as a critical measurement on your "dashboard". You set a target of keeping inventory under $100,000 and establish a bonus structure around either succeeding or failing to reach the target. In other words, you will pay your people more to hit the target. The theory states that your metric, inventory dollars, will become manipulated or unreliable.

Anyone who has been around management for awhile has probably seen this phenomena. Just before the measurement is scheduled to be taken (usually at the end of the fiscal year) the inventory drops. Production is modified, items are shipped early, or sometimes materials are even stored off-site. The point here is that because the inventory metric was emphasized and a reward system was created around it, the inventory metric became a poor indicator of performance. Nothing really improved.

Look at any company that has quarterly sales targets. If you were to plot the day to day sales by quarter, you would see a pattern emerge. When the end of a quarter comes, the sales person makes one of two decisions. If the person is going to hit their sales goals for the quarter, they will choose to delay sales near the end of the quarter for a few days or even weeks to help with the next quarter. If they are short, they will go out and try to close a bunch of sales quickly at the end of the quarter, even if some of the sales fall through later or even if the pricing is not as good as could be. Working as an engineer purchasing equipment, I have seen this phenomena many times where a company offers a discount if I purchase the item by the end of the month.

Does this really improve the process? Does this really add value to the customer? Think about the Data Uncertainty Principle and how it applies to your organization!