Dumb Stuff to Avoid: Paying for Forecast Accuracy Is a Lot More Costly Than You Think
In my last blog, I began explaining the changes in target-setting when switching to a rolling forecast to eliminate annual budgets. Target-setting switches to midterm targets that are relative to peers, competition, and world-class benchmarks.
This frees rolling forecasts, which should be used to help you steer your organization. But you need to avoid the devastating practice of “paying for forecast accuracy.” While many have started to advocate this, it is one of the dumbest things management can do. The cost payment is magnified by the tendency to suboptimize results.
The only time a forecast could always be accurate is if none of the surrounding processes had any variability. The reality of today’s complex world is that we are surrounded by variabilities often interacting with each other. We are wise to measure and understand this, but paying people to hit certain forecasting numbers is a prime case of what quality guru Dr. W. Edwards Deming called “tampering with the system.” If management tampers with it, you never get an accurate, unbiased measurement. As a result, you never understand the system’s variability, nor can you improve the process to reduce this variability. So any improvements are illusionary and unsustainable.
When you think of forecasting, you often deliberately act to negate the forecast. If the navigator forecasts that the ship will hit the rocks, the captain immediately acts to achieve a different result. Likewise, if your forecast informs management that the current actions will only produce 4 percent growth and your target is 10 percent, you would want to plan additional corrective actions to move closer to the target.
None of this should be tied to incentive compensation. Why not? Because people should be paid for what they actually achieve, not for what they hope or promise to achieve. If you separate the two, you get a much more honest discussion of what can be achieved, what investment and risks are required, and what other alternatives could be considered. You want real results that put an end to tampering.
During the week of August 31st, I will be in eastern Pennsylvania, New Jersey, and New York City. I look forward to your questions. ###








