A Better Performance Management System (Part 2)

In September 2021 part 1 of this blog was published (https://www.techoregon.org/blog/a-better-performance-management-system/.  We asked that you stay tuned for future blogs that will cover the retention data from using this merit review system, tweaks to move undesired attrition to the lower value quartiles, how to handle employee challenges to the value rankings, and how to conduct the merit review (and promotion) conversation.  We will cover the first two items in this part, and also address a concern raised by a reader of the initial blog article.For a brief review, all employees were grouped by title/pay grade, and quartiled in each group by salary and value to the company.  We then calculated the difference between their salary quartile and value quartile.  As an example, for a person in the bottom quartile for salary (4, lowest paid) and top quartile for value (1, most valuable), this person would have a difference (delta) of 3, indicating that they are significantly underpaid.  A negative delta would indicate that someone was overpaid.   Assuming that the company has determined that they will grant an average increase of 4.5%, the below table is an example of ranges for raises based on the deltas.  The remainder of this blog will cover our actual experience in using this approach, and tweaks made to reduce undesired attrition the following year.Salary Quartile (across)Value Quartile (down)123414-5%5-7%7-9%10-14%22-4%4-5%5-7%6-8%30%2-4%4-5%5-7%40%0%2-4%4-5%(Note: The values in the cells are the annual salary increases.  Since the difference is a zero-sum calculation, where positive and negative differences will exactly balance, the average increase across all the cells in the above table is the target 4.5%.)How did this new merit review system work?  When we implemented this system, attrition for engineering staff at our company and in the metro area was ~20% annually.  After implementation in a  group of 600 engineers, attrition dropped to 7% annually.  It remained at ~20% in the other engineering groups at our site, and in the metro area.  It was so successful at reducing attrition that corporate HR sent someone to chastise our team for not following the 4.5% annual raise guideline.  Once corporate HR understood the system, they commended us.  However, we continued to study the attrition, and found that nearly all of the 7% annual attrition was from people in value quartiles 1 and 2 (most valuable).  Around the same time, an employee approached me (taking advantage of my open door policy as a VP) to voice her/his displeasure with the new system.  S/he thought it was an improvement, but would we not prefer to retain a 1/1 over a 3/3, both with a delta of 0?  Doh!  Our management team agreed and we tweaked the table as shown below.  Rather than basing the merit increase range solely on the delta, it was skewed in favor of those in the higher value quartiles.  The average increase across the table maintained the desired 4.5% average merit increase.Salary Quartile (across)Value Quartile (down)123415-7%6-8%9-11%10-14%23-5%4-5%5-7%6-8%30%2-4%3-4%4-6%40%0%0-2%2.5-3.5%How did the tweaked system work?  The next year, attrition remained at 7% annually, while the other engineering groups at our site and metro area remained at ~20%.  Over 80% of our attrition was now from people ranked in the bottom two value quartiles.  Attrition from the top two value quartiles was just over 1%.  It would have been interesting to track the system over subsequent years through changing economic and competitive conditions.  Unfortunately, corporate management called corporate HR onto the carpet.  Corporate management did not want one group using this revised system, as its implementation made all other engineering groups look bad.  HR did not believe that any other engineering groups would put effort into running such a system.  It was an unfortunate decision, as we were never asked about our experience and found the system required little additional effort.  It also had the benefits of manager team building across groups, and encouraged a lot of productive conversation among managers on the strengths and weaknesses of their engineers as viewed by other managers with teams that worked together.  We also want to address a comment received after the initial blog was posted.  A reader pointed out that this merit review system would not do anything to reduce attrition for a small company that could not pay competitive wages with larger technology companies (e.g., Microsoft, Amazon, Google).  This comment is true.  A company has to ensure that their compensation ranges are competitive with other employers.  However, our experience showed that having competitive ranges did little to reduce attrition from the two higher  value quartiles.  Recognition and reward for people in those quartiles, and particularly rewarding those who were most underpaid, significantly reduced attrition. One unintended benefit of this salary/value system.  When business conditions required a workforce reduction, we were able to identify the most overpaid as people for reduction, rather than the least senior or highest paid.  Is that not how every company should look at this problem?  Please reach out to me at markneuhausen@gmail.com if you have questions or desire assistance in implementing a system such as this.  As a TAO Executive in Residence (EIR), my assistance is included in your company’s TAO membership.