Wow. That Data is Terrible!
Almost every compensation professional I know can guess the base pay of mid-level accountant or software engineer with near-perfect accuracy. Given regional variations and other factors, most guesses are within 5% of current market data. This consistency and predictability is a foundation of trust when we are pricing jobs that are less familiar or more volatile. If only it was always so easy.
Annual incentive plans add complexity and noise to pay data. Even when you have confidence in the target and actual levels in the survey data it takes real confidence (or ignorance) to make decisions on how the plan pays out in good and bad scenarios. Even with this uncertainty, most compensation professionals do a great job and getting these numbers within a reasonably accurate range.
Long-term incentives are a different animal altogether. In the best cases, the data is noisy or opaque. A good example of this is the pay packages for executives at publicly traded companies. The data seems like it is all there if you know how to dig through it. Once you get into the thick of things a talented executive compensation professional can make sense of it and provide real guidance. When you move from publicly traded to privately held companies things can go haywire.
I recently looked at two industry-leading data sets. They were for the same size companies. For the same industry. For the same region of the U.S. But that’s where the similarities ended. As I mentioned earlier, a good compensation professional can usually guess the base pay amount for a standard position within 5% of the market. Imagine thinking someone should be paid $100,000 and looking into the data to see they should be paid $853,000. It would probably make you question your abilities or the data. The differences in the two sets of equity data I was reviewing ranged from less than 5% to more than 800%!
I spoke to experts from both survey providers. Each expressed their absolute confidence in their data collection and analysis processes. Most disturbingly was neither thought that the differences were of great concern. Now, I work on equity compensation programs for all sorts of companies nearly every day and have done so for more than twenty years, but this nonchalance in the face of craziness reminded me how hard this can be for someone who only deals with this data once or twice a year.
In some cases, ignorance is bliss. If you had only ever seen one of the sets of data I reviewed, you may not have any idea how different it was from the other. Heaven forbid you saw both and decided to verify which was right by purchasing a third set. I can almost guarantee that the third set would raise just as many questions. We love to say our industry is a combination of art and science, but I think most of us were thinking of art a bit less like that of Jackson Pollack.
I love to wrap each post up with some type of solution, so here goes… When dealing with equity, especially at private companies, expect the data to be terrible. Get more than one set of data so you can have a reasonable range for the terribleness. Then do your process in reverse. Instead of looking at the market data and determining how to fit people in, look at your people and determine what you are trying to accomplish. Then determine what levels of equity and types of features will help you reach your goal. Finally, take a look at the market data and determine if you are, based on your needs, within the right range of accuracy.
I find this type of work to be the most fun and interesting part of my job, but I respect those who want a more solid foundation. I’d love to hear how you deal with this challenge (or not).