Bill Kutik hates the merit matrix and I suspect he is not alone.
The merit matrix is not dead -- not yet, with recent WorldatWork research confirming that about three-quarters of surveyed member organizations (which tend to be larger, more well-established employers) have a merit matrix in place. But should it be? Will it be?
In his recent HR Technology column at HREOnline, Death to the Merit-Pay-Increase Matrix, Kutik (co-chair emeritus of the annual HR Technology Conference and Exhibition), shared his issues with the matrix:
As you likely know, it is the paper form (spreadsheet or compensation application) that companies use to control growth in wages (except for those at the top, of course) and for HR to fulfill its traditional role of treating everyone the same. Ironically, often without regard to merit!
I'm not here to throw down a blind defense of the merit matrix. It is not a perfect tool. But like many tools, it simply does what it is designed to do by those who use it. And so we have to distinguish here between the sins of the tool and the sins of the user.
Although there are many versions and enhancements, the basic merit matrix is designed to guide the determination of a merit-based salary increase, considering two factors:
- The performance level of the employees, and
- Where the employee is currently paid relative to the mid- or control point of their salary range, which is a proxy for the going market-rate for the job they hold.
Basing a decision about whether and how much to raise someone's salary on how well they're performing and how competitively they're currently paid -- while neither perfect nor comprehensive in a total talent management sense -- isn't a terrible place to begin, tool-wise. The challenge comes in how the tool is being used, the quality of the information (performance, competitive rate) on which it hangs, the thought and strategy behind its formulation and (ultimately) what goes into the little matrix boxes.
There is nothing embedded in the merit matrix that forces you to use it to hold down growth in wages, treat employees the same regardless of merit or drive high performing, highly paid employees back to the midpoint. Those are talent investment funding and tool design decisions, Bucky. Too many merit matrices have been running along on auto-pilot for too many years. That's bad HR, and no tool is going to produce good results under the direction and influence of bad HR.
Having said all that, a tool as basic and two-dimensional as the merit matrix certainly has its limitations and the increasing power of data and analytics, fueled by technology, ought to bring exciting new possibilities to the arena of salary decisions. In his column, Kutik shares news of a new solution being debuted at the HR Technology conference by Saba. The solution uses a "new Big Data and machine-intelligence platform" which analyzes a number of data points, both internal (HRMS) and external (pay surveys, etc.), runs them through its proprietary algorithms and develops a custom-tailored rcommendation on how much more each employee should receive.
While technology-powered Big Data has much to offer us in terms of understanding how compensation influences and drives work outcomes and employee choices, I think we need to be judicious about delegating compensation decisions. Integrating machine learning into the arena of pay decisions must be done with a full appreciation of growing pressure building to bring clarity and transparency to those decisions. Can you imagine responding to an employee's question about the amount of their base salary increase with something like "well, we've got this algorithm..."?"
Bottom line, I'm not sure we should write off the merit matrix just yet. But evidence abounds that we may have a lot of work to do if we want to bring credibility back to the tool.
Image "Money with Magnifying Glass" courtesy of Sujin Jetkasettakorn/FreeDigitalPhotos.net
Bill has been an IT guy for his career and his talents might be better suited for getting the bugs out of the latest HR software. His criticisms of the matrix were old news and his work-arounds are standard practice.
Posted by: Todd the IT guy | October 06, 2014 at 12:25 PM
Todd:
Bill may indeed be an IT guy - but the criticisms he shares about the merit matrix are not his alone. My experience would suggest that many managers and employees share many of these same concerns. That's why I thought it was a good topic to address - particularly in light of the data-and-analytics alternative he promotes.
Thanks for sharing your point of view!
Posted by: Ann Bares | October 06, 2014 at 08:26 PM
Wonder if his "analytics" approach employs the forced distribution method? Methinks it would be extremely unlikely that any IT solution today would be better than the spreadsheet versions of the distant past that analyzed combined weighted survey inputs, derived a normative MRP, input the job value for a configurable range midpoint best fit, then checked performance ratings and their relative weights to compute the proper personal increase percentages to apply to the appropriate midpoints to perfectly allocate a fixed increase/bonus budget pool. No forced performance distributions. No slippage, either.
Posted by: E. James (Jim) Brennan | October 08, 2014 at 04:59 PM