In this week's Business Week (I know enough already, but it was a great issue) in The WelchWay section (Jack and Suzy Welch) they answer a reader's question on ranking employees into performance categories of the top 20%, middle 70%, and bottom 10%. It shouldn't surprise you that the Welchs backed up the rankings and you know what so do I.
A while back at AT&T, we used to have the classic annual review where the district managers got together and tried to get as many people into the best rating. Of course it was utter chaos as pretty much any breathing body would get a good rating and what the discussions came down to was a beauty contest and good old fashioned horse trading. Then finally, AT&T handed down a 16 box grid with the x Axis being how you performed your job and the y axis being how you did versus your annual objectives. With the 16 box grid came a forced curve with which percentages the population had to fall in. As I recall only 5% could only be in the best square (most money and a vote into True Spirit), 15% into the other excellent grids, 10% into the bottom, and the remaining 70% spread around in the middle.
It was that year when a manager walked into a performance review and said one of the dumber things I can remember "I can't stand this force fit. I think everyone at AT&T is above average so everyone should be in the best square." Obviously, she never took a statistics class and looking back with the demise of the old and better AT&T, was a little off-base. It shouldn't surprise you readers who snapped at her and told her in front of the rest of the group what a moronic statement that was and not everyone can be at the top. It doesn't work that way in the real world or even in professional sports ;-) and it certainly shouldn't work in an AT&T performance review. This manager's team did fine in the review, but she fit them into a curve, and pretty much hated me from that day forward. Oh well.
There were two times the curve didn't work out for me. One was for a manager of mine who I had rated higher than someone else in my group, but the rest of my district peers liked another person in my group. They flipped my ratings around moving my lower ranked person from the top 15% grid into the top 5% and lowering the other manager. That was a tough conversation, but the affected manager made it into the top 5% a year later.
The other time was how I got screwed out of a True Spirit trip and a top 5% ranking for building AT&T's Personal Network offer. That product was dead in the water until I got there, as other managers played around with our tech and product teams by over-promising on start dates, playing around with included products, and price points. We got the offer launched on the Super Bowl and all I got was a pat on the back as I lost out to someone in a popularity contest. (It doesn't happen often when I lose a popularity contest).
Force curves are the only way to reward people for an outstanding job. If you don't have them, then everyone will end up in a great category, getting similar bonuses that reward people for mediocrity. It would be like all of Major League Baseball's players making the All-Star team which would be boring to watch and probably end up in a tie.