Can organizations use attrition rates to improve manager performance? It’s tough to do.
In an article last week, I started a series on how managers receive feedback on their own performance. Most managers don’t actively solicit performance feedback, let alone from their employees, and as a result, the default moment that a manager receives feedback as a manager is when someone quits. As performance feedback, it fails. But what about when a bunch of people quit over a period of time? That has to say something about a manager, doesn’t it?
It does! But there are problems, and as a design for getting information to the manager so that they can change their behaviors, it fails. Here’s why:
Attrition rates = slow, lagging data: Many companies’ HR departments will track attrition rates for managers or divisions within the company. They may even divide attrition between “good attrition” (lower performing or problematic employees who leave) and “bad attrition” (high performing employees who leave). This is excellent information, since it can give visibility to the HR department and the management team where the best and worst rates are in the organization. They can then do a study to identify what is being done in the “high good attrition/low bad attrition” areas, and vice versa. Then they can identify the teams and organizations that are having the most problems, and then work on finding ways to change the behaviors, and try to implement a program that improves the areas where the bad attrition is the highest and good attrition is the lowest. Then they can measure the new “good and bad” attrition rates and see if there has been improvement. . .
When looking at this, you can see that, as a design, it is very, very, very slow in identifying where the poor managers are, and perhaps not even who the poor managers are. It is even slower in identifying what it is the managers were doing that was so bad, and yet slower even still in addressing it with the managers who are causing the problems. And even slower yet at identifying whether the effort made a difference in the organization’s attrition rates.
This is a long time to take to identify and do something about managers who are actually causing good employees to leave.
Attrition rates: Do they find the bad managers? And if there was a really bad manager in an organization that was contributing to that high “bad attrition” rate, that manager could be long gone (perhaps to better opportunities?) by the time it is addressed, and the manager improvement program risks being targeted at bunch of people who had nothing to do with the bad attrition rate.
Attrition rates: Not performance feedback: As performance feedback, attrition information is neither specific nor immediate, which is a basic element for artful performance feedback. In fact, the performance feedback is vague and highly delayed. So what you get is a representative from (most likely) the Human Resources (HR) department approaching the department with the high “bad attrition,” and informing the department heads that they have bad attrition, and (perhaps) HR has a program or willingness to help to fix it. So whatever program comes into place to resolve an attrition problem falls essentially to the problem of “public feedback”, since the intrepid HR department is saying, “this is a general problem that needs to be addressed by all the managers in this department.” The problem with public feedback is that it is ineffective at changing bad behaviors, perhaps exacerbates bad behaviors, and even dampens good behaviors.
Attrition rates: Organizational resistance: Most likely, the division with the high “bad attrition” will not do a detailed effort to identify the best and worst managers in the organization. That division is more likely to respond with resistances such as, “We don’t have bad attrition,” “We’re addressing it already,” or “That was a long time ago.” A more likely discussion is that HR will instead target the head of the department with the lowest scores and ask the head of the department to change things. A few problems with this approach: That department head may neither be capable nor interested in improving in this metric (both highly likely), and with the delay in having meaningful metrics, the department head may claim that it is not relevant to her performance, but the prior department head’s performance.
Ergo, slow and ineffective: The basic problem with tracking attrition rates is that it is the ultimate lagging indicator – it lags even more than sales data, market share info, or profitability, since it would likely be measured over even a longer term than the finance calendar. Attrition rates provide for insightful information about the overall health of an organization, and can have a great impact on a new overall strategy going forward from the top levels of the organization (“We want to make this a great place to work – we’re reviewing our compensation program. . .”) However, for providing performance feedback to managers and changing their behaviors, it is an ineffective method. You slowly get the insights into the reasons people leave, you have to wait until you amass enough statistics, and it only provides meaningful information about organizations large enough for statistics to have significance. Individual managers with poor habits are rarely identified, nor are they given specifics about what to do differently.
So attrition rates do provide information, and the information does come from the employees, but it does not constitute feedback. A manager who may be responsible for the poor attrition rates can easily hide in this situation, and the efforts to solve it are typically not highly targeted. This may explain the overall high cost of low quality management.
In my next article in this series, I’ll explore the management survey and 360 degree feedback as a feedback source to managers.
What’s your opinion? Does your organization look at attrition? Have you ever been in a discussion about attrition rates? What was done to address it? Did the results of your discussion ever get to the managers who were least effective?