Maybe some of you have lived through a misguided company’s application of talent metrics. The stories that many of us have show how these metrics can be incredible de-motivators.
One example comes to mind of a company that decided to develop a personal ‘earned value’ measurement on every employee in an IT organization. After creating individual scores the management ranked everyone, published the results on a website, and on a screen in the company lobby. These metrics were also considered in pay treatment … for a short while. New management came in and considered the idea nuts, and then deimplemented it.
Every poor performer figured out how to game the system. Every talented performer either ignored the metric plan in favor of getting work done, or they left the company.
Let’s consider some ideas that were posted in a recent Harvard Business Review article, “Competing on Talent Analytics’ by Davenport, Harris, and Shapiro. Please read on.
Companies that use analytics for employee management can create tangible value for themselves as long as they avoid these mistakes:
1. Making analytics an excuse to treat human beings like interchangeable widgets
2. Keeping a metric live even when it has no clear business reason for being
3. Relying on just a few metrics to evaluate employee performance, so smart employees can game the system
4. Insisting on 100% accurate date before an analysis is accepted–which amounts to never making a decision
5. Assessing employees only on simple measures such as grades and test scores, which often fail to accurately predict success
6. Using analytics to hire lower-level people but not when assessing senior management
7. Failing to monitor changes in organizational priorities, thus creating irrelevant-if accurate-analyses
8. Ignoring aspects of performance that can’t easily be translated into quantitative measures
9. Analyzing HR efficiency metrics only, while failing to address the impact of talent management on business performance
What are these guys missing? What would you add… or challenge?
Chime in, please.