Evaluations. Performance conversations. Employee reviews. Whatever they are called, this month is a time where millions of employees are assessed for what they’ve done, with paperwork and meetings attached.
In many workplaces, they are formalities, and sometimes actual conversations that set a direction for the coming year (or quarter). But in some companies, such as the Jack Welch-era General Electric, Microsoft until recently and now Yahoo, evaluations take some form of ranking, with real consequences for those unlucky enough to be tagged deficient.
So-called “stack ranking” has its defenders, and avoiding such a system still leaves the questions “How can I assess how my employees are doing? How can they know how they are doing?”
These are questions I ask myself, as a manager and someone who deals with leadership issues on a day-to-day basis. I also reflexively abhor paperwork and the stiffness of formal performance conversations, yet I understand that whatever the antidote, winging it isn’t the solution. So, to answer these questions, starting with the problems of today’s appraisal systems and then discussing what ought to be done, I talked with Aubrey Daniels, a noted expert and author on performance management and other management, leadership and workplace issues, as well as an occasional SmartBlogs contributor.
He also recently started The Aubrey Daniels Institute, which is “dedicated to increasing the understanding and advancing the use of the science of behavior (behavior analysis) in the workplace.” Below are some key questions and answers as informed by our conversation.
What do we mean by stack ranking, where did it come from, and what’s the problem with it?
Essentially, stack ranking is any system that involves using a curve to sort performance from the top employees to the bottom, and was popularized by Welch in the 1980s. Welch’s system included firing the bottom performers because, as Daniels puts it, “most supervisors/managers don’t have the guts, as he might say, to fire the bottom 10%. … so we have to have a system to force them to make a decision.”
Defenders may say it’s a way to help employees do better, but Daniels disputes that. “One problem is, of course, is you don’t hire on a ‘normal curve.’ You don’t say, ‘Well, we hired too many smart people, we need to hire some dumb people.’ So, the statistical assumption is wrong anyway.”
“The more troubling part is the fact that it puts people in categories and gives them a label. … once you put people in a category or give them a label, you treat them like the label. And so they become the label.” This is bad for the employee but also bad for any manager hoping to evaluate objectively. “If you think someone is a poor performer, you’re going to look for instances to validate that category,” Daniels says.
But can’t it be helpful to have some categorization?
Categories are nice for sorting, sure, but by their nature they divide co-workers, argues Daniels, and in theory and according to data, the notion that performance appraisals help employees become better doesn’t hold up. Even the well-meaning idea of “high performers” necessarily means that some people are “low performers,” creating a noticable separation among the workforce.
The problems of appraisals of that kind include, as Daniels describes it:
- Documenting employee performance doesn’t always have the intended effect. These appraisals can be used by employees to defend themselves and may favor them, depending on the interpretation.
- The appraisal process “is so 20th century.” Such appraisals, Daniels says, have a history and reputation of not working, being disliked and not backed up by data.
- Frequency isn’t the problem. “If it’s a bad process, doing it four times a year doesn’t make it any better than doing it once a year.”
- Lack of objectivity. “Some of the numbers are pretty clear — production numbers or whatever — but there’s always an opportunity to fudge the ranking in the end,” Daniels says. Rankings can be altered based on whims of managers, whether to save face, to punish a disliked employee or excuse or reward a well-liked employee, even if unconsciously. Certain qualities that are part of rankings cannot be quantified easily, Daniels says: A manager may be able to say someone demonstrates leadership, but how can that be objectively measured? Maybe it can, but most processes aren’t equipped to do so.
- Managers circumvent the system if they consider it unfair. Daniels talked about a manager of 30 nuclear engineers who was ordered to stack rank, where at least a couple of engineers would be marked deficient each evaluation period. However, those employees wouldn’t be fired unless they received two consecutive such rankings. So, the manager told his team that some people would be ranked deficient on occasion, but never twice in a row. Problem solved, except for the larger cultural problem that drove the manager to such action.
- Young people don’t like labels. The much-talked-about culture shift toward “everybody gets a trophy” leads to younger people in the workforce “not only resent their label, but they also resent other people being labeled,” Daniels said. That collaborative spirit goes against the separation and calling out that labels bring.
So what’s the point of performance appraisals? Who is responsible, and for what?
The key question in Daniels’ mind is this: Is the manager doing the utmost to help employees do the best job possible? Seen that way, “appraisals are really an appraisal of the manager, not the employee.”
“A supervisor in a business should be basically trying to find ways to increase the productivity, the quality, the personal attributes” of the team, Daniels says. Step one is to back off: Managers, generally, employ people who know more about the technical aspects of the operation, he says. They should be going to those people for questions. The supervisor’s expertise lies in helping with “matters of human behavior” and coaching and encouragement, not in micromanaging the details, particularly if the supervisor’s knowledge isn’t up to date.
When appraisals are manager-focused, then the focus is on making all team members as good as they can be, not separating out the weak links. As Daniels shared, when a team is asked to lend a member to another team or department, managers will often pick the poor performer, as they don’t want to lose their best people, and if the transferred employee fails, so be it. But in a system where managers are held to account, dumping the weak link will backfire because a manager who loans out an employee isn’t dumping them on the new department — that manager retains a stake in that employee’s success or failure throughout his or her journey.
What if the manager is out of line — being vindictive, petty or simply incompetent?
Who is the manager of that manager, Daniels would ask. It would seem to me that this is a major concern for companies looking to decentralize or de-formalize performance appraisals, particularly at very large companies, where a rogue manager can be invisible, or very small companies with few layers, where one bad supervisor can control nearly everything, or be his or her own boss. Suffice it to say, however, that a company that does not hold managers accountable will find this to be a much bigger problem.
So, old-style appraisals aren’t working. What’s the new plan? Is there a “plan”?
In short, communicate in a streamlined fashion. Daniels goes so far as to say of appraisals, “Eliminate them.” In their place, a less rigid, paperwork-free but more personal approach:
“You look at the individuals in terms of, OK, what do you need to change about your behavior and how can I help you?”
Not having performance appraisals as commonly known doesn’t mean no conversation. Managers absolutely should sit down with employees, Daniels says
“Let’s think about what’s going on this year, and think about things you think you need help with. Let me tell you a couple of things I’ve seen that I think may be helpful, and let’s see what we can do about it. I mean, why do you need to do any more than that?”
Take time for reflection, planning, feedback, but formality isn’t a requirement, Daniels says — engagement is. Are there problems? Are they fixable? How can they be addressed through the difficult-but-necessary task of positive reinforcement? “If the supervisor is not positively reinforcing the people who work for them, then you’ve got a problem already,” Daniels says.
Worth noting is that the best path to strong performance management begins well before any appraisal — hiring and promoting well. As a “first screen,” Daniels says, find candidates with good people skills, who are approachable, friendly, looking to help rather than fixate on failure. Promote those people if they are already in the organization — and don’t simply reflexively promote the highest performers and assume they are ready and properly prepared to lead.
In short, ask, “How do people work best?” and help them get there. It’s not easy, requires a lot of work, and is a philosophy based on people rather than process — a scary proposition, especially for command-and-control structures. But it may be a proposition worth exploring at your company.