Can a forced to a normal distribution annual review system work?

If we believe we do not hire to a normal distribution, can a  ‘forced to a normal distribution’ annual review system be effective? If not, what are the alternatives? – Will Goss, R&D Manager, Champion Technologies, Inc.

Community Responses

Response 1
The simple is answer is “yes.”  But as usual, it’s not that simple.

If a company truly rewards A players and eliminates C players, then over time the company should comprise better than average employees as compared to the available market.  Better than average players should drive better than average internal systems, better than average corporate performance, and better than average growth and profit.  Therefore, a bell distribution of rewards tied to corporate profit and other sensible internal measures still corresponds to a pay-for-performance scheme with better performers getting a higher share of rewards.  Acceptably good performers whose performance is lower on the company’s bell curve can still outperform the market because in a well performing company, a smaller share of greater profit is better than market average.

As to the question of what is the right distribution to use, I have a very difficult time believing that it will be substantially non-normal owing to the central limit theorem – the averages of even highly non-normal distributions accumulate as a normal distribution for all but the smallest companies unless applied to small numbers of employees that are highly compartmentalized and independently measured – say groups of five or fewer employees.  On the other hand, one could make a case that for a collection of very rare events (i.e., breakthroughs and disasters) the limiting distribution tends toward a Poisson distribution, but instead of worrying about the actual distribution of rare actors, I think it best to handle super-performers (or super-disastrous employees) by making infrequent exceptions to the system.

In my experience, the biggest problems with employee reward systems are the fear managers have in dismissing mediocre employees and the contamination of corporate performance measures with external measures such as market indexes.  Both issues ultimately reduce the correlation between pay and performance and conspire to underpay top performers and overpay under-performers.

– Joseph Colannino, Director, R&D, John Zink Company, LLC

Response 2
My experience is that forced distributions create unnatural competition among team members who immediately know that 10% of the team will be laid off or put on performance improvement, even if all team members are performing well.  The failure of forced distribution is that in small teams or high performing organizations, it poisons the informal and formal collaboration necessary for innovation and advancement of the business.  Proponents will say you never actually need all the people you have.  Most middle and lower managers will disagree.  Proponents will say competition for position and rewards brings out the best in people.  Most middle and lower managers will agree if the WHOLE team is involved and participating for a whole team reward or outcome.  If the effort involves a subset of the team, then unhealthy positioning and personal agendas will be the order of the day.

– Art Paton, Motorola

Response 3
I believe that a forced normalization is necessary for large organizations, where each group evaluates and scores their employees separately and those scores are then used (and thus need to be normalized) at the organizational level in order to allocate raises and bonuses equally. I believe normalization to an average is better than normalization to a distribution. Forced distribution leads to unnatural allocations at both ends of the spectrum. Special care also needs to be taken with small groups or after layoffs if forced distribution is used. Regarding the question about “normal” distribution, that should be a concept defined for each organization with respect to the organization’s expectations; the expectations may be higher than on average outside the organization, but within the organization, one would still expect a distribution for employees that are above, at, or below those expectations.

– Markus Fromherz, Chief Innovation Officer, Healthcare, ACS, a Xerox Company

Response 4
Although we do not practice a strict forced distribution per se, we do encourage a strong differentiation amoungst the workforce as part of our annual review system and with the right management support and emphasis we are regularly able to meet distribution targets.  Of course no one intentionally hires to distribution; your intent when hiring is to hire the best, the best fit, the best value.  But as part of executing, the chining bar always goes up and people’s performance is always a variable.  Too often the focus of this discussion tends to be on the lower end, the lower performers and how this impacts them.  A strong focus on the other end, the higher performers sends a more positive message to the team.  It is critical to differentiate, recognize and reward those who are contributing significant results to the organization if you want them to stick around and continue to do so.  This also sends the appropriate message to the rest of the team as well and this general message of trying to ensure we reward those who have contributed heavily is generally received well by all.

– John Gibbons, R&D Finance, The Boeing Company

Response 5
Yes it can work.  Kimberly-Clark has been doing this for quite a few years.  It puts a system in place to make sure the managers make the needed tough decisions and have the heart to heart conversations.  One of the alternatives would be forced ranking.  We did that for some time in the past.

– Yong Li, Global Director, ECB Sciences – Corp R&E, Kimberly-Clark Corporation