People Analytics In Compensation
Compensation Analytics: the opportunities and challenges of using data to determine pay
The Article Here: https://www.visier.com/clarity/compensation-people-analytics/
The above article focuses on three issues (counter offer negotiations, identifying discriminatory pay practices, and justifying compensation increases) related to compensation in organizations, and how people analytics can help solve these issues. While I agree that data should be used in each of these scenarios, analytics are by no means a cure-all if they are used thoughtlessly. I’ll address each point in turn, highlighting the articles point of view, and then agreeing with or challenging the asserted views.
Counter-offer negotiations:
The gist of the article above is that a manager naturally wants to match an offer that their employee received from a competing organization, and that data should override that gut instinct to either justify matching, or confirm that your own org should not be willing to match. The article highlights data points including performance ratings, performance incentives achieved, and other attributes such as tenure, level, and department for consideration. While I completely agree that any manager should consider the going market rate for their employee based on the above criteria, they must also consider attributes that are much harder to define numerically, like personal and team chemistry, institutional knowledge and connectivity, and whether or not you (the manager) could easily backfill the position. I think data should help inform the counter offer, but should yield to the “softer” traits that only a manger might pick up on. We all know people we have to have on our team, and the X factor that makes us feel that way is often hard to quantify.
Identify Discriminatory Pay Practices:
I believe this is a great purpose for people and compensation analytics within a company. The article outlines cutting overall compensation levels and compensation increases by age, gender, ethnicity, tenure, education level, etc. in order to identify areas of the company where you (management) may unknowingly pay too much or too little. In a large organization, this could work. Assuming you can look internally and control for such factors as level at the company, tenure, and department, if there are gaps in pay among different ethnicities or genders, those should be surfaced. A few challenges here: 1) data collection could be difficult, and at a small organization that data is thin even if employees are willing to provide information on ethnicity, religion, gender etc. 2) a separate challenge for small organizations is the need to look to other companies for comparables (given the thinness of data), which then hinges on those outside companies practicing non-discriminatory compensations schemes. Again, reducing unknown compensation discrepancies seems to be a great use for people analytics and data, but organizational data may not suffice.
Justifying Compensation Increases:
The article above posits that managers should use data to justify pay increases, and to better target those increases based on department and level, depending on what parts of the organization need a boost in retention rate or added performance incentives. Again, I agree with the idea here but the issue is that pure performance data, or data that captures the threat of external poaching, could be hiding inherent bias. Employees in high demand may be in high demand because of personal connections at external companies, or they may be outperforming peers because they are better connected to senior leadership or exhibit a similar communication style to their manager, or any number of reasons that are not directly related to output quality. Using raw data to justify pay increases for these employees may in fact lead to the very problems we tried to solve above in terms of pay equity.
In short, I fully support the use of data in compensation decisions, but I also acknowledge that using data analysis to try to eliminate bias may in some cases actually perpetuate that same bias. Looking forward to hearing thoughts, and I hope to check out any similar articles people read!