Welcome back. In this video we're talking about designing performance evaluations. So designing performance evaluations to do what? First of all, we want to design evaluations that can develop employees, get accurate data that can be used for things like compensation, promotions, and so on, and also develop evaluations that avoid conflicts and avoid bias. And how are we going to do that? The two main steps are going to be good benchmarks and good communications. And the first thing we're going to look at are benchmarking inputs. When we're talking about benchmarking inputs, we're talking about benchmarking employee behaviors. And so for here we're going to introduce a new term. This is a behaviorally-anchored rating scale, or BARS for short. So BARS is a scale that maps behaviors onto a specific performance rating scale. So let's see how that works. So let's take an example of a BARS for customer service. So here you'll notice that there are five different scores, a rating of 5, 4, 3, 2, or 1. And each of those scores are tied, or they're anchored, to a very specific set of behaviors. That is, to get a rating of 5 on customer service, this is a person who must discover, probe, and solve problems that customers don't even know about. And to get a score of 1, you would have to have little attempt to understand customer problems. And so what are we doing here? By tying each score to something that seems like something more of an observable, objective behavior, we're trying to make it so it's more easy for a manager to justify a score that corresponds to a certain behavior. And also something that could be more objective and more comparable across perhaps different managers who might have different biases or might have different tendencies to rate people differently. That is, regardless of where you are in the organization or who's doing the ratings, you should expect that someone who's getting a 5 is also someone who solves problems that their customers might not even know about. So when do we actually use these behaviorally-anchored rating scales? We're typically talking about evaluating performance based on these behavioral inputs. If we know what employees should do to be successful, we can actually measure what it is that employees do and that usually involves jobs where employees have more supervision. That is, to actually implement BARS, we'd have to have a manager who's actually in a position to observe how that employee interacts with their customers. Are they able to engage in kind of a problem-solving conversation? Do they make efforts to learn about their customers' problems or so on? And so some of these behaviors might be very difficult to actually measure based on their outputs, because different clients could defer, they might not collect data on the outputs, even though you can observe what the behaviors those employees are doing, or so on. And so again, if we can observe those behaviors, then we might be thinking about benchmarking on inputs, that is, using something like a behaviorally-anchored rating scale. Next we'll talk about benchmarking on outputs. When we're talking about benchmarking on outputs, we're not talking about trying to measure what it is that employees are doing or their behaviors. We're talking about measuring what it is that employees actually contribute, what it is that they produce, what they output. And one of the methods for this is a management by objective, and so that's a method for benchmarking employees' ratings to their achievement of prespecified goals. And when we're talking about management by objectives, an example could be, again, a 5-point rating scale. So here, to get a rating of 4, the objective would be to meet 100% of the new business target. And to get the highest ratings, you'd actually have to beat your target objective. And at lower levels of meeting that new business target, you would get a lower level. And so how is this different from a behaviorally-anchored rating scale? Well, you'll notice that, again, you're not measuring their behaviors, not measuring what it is they try to do or their effort or whatever they're trying to do to get new business. You're only measuring whether an employee actually achieves this outcome or this objective. And so typically we're talking about measuring based on outputs, or using a technique such as management by objectives, if we know what employees should accomplish to be successful, not what behaviors they should use in order to be successful. We're also typically going to benchmark on outputs and use techniques such as MBOs when we can measure what employees accomplished. That is, for example, if we're talking about an airline flight attendant, then it might be very easy to see that an employee's behaving courteously, but we might not be actually measuring outputs such as customer satisfaction. And in that case, we'd want to use something like BARS as opposed to management by objective. But if we're talking about a job such as salespeople who would have a lot of autonomy, whose outcomes would be very easy to measure even if we aren't necessarily directly supervising them. Well, in this case, since we can measure those outputs, we might be more likely to use a technique such as management by objective. And so how do we design our MBOs? Well, we can typically use the acronym SMART to help us design them. That is, management by objectives, those objectives should be specific, measurable, attainable, realistic, and time-bound. So for example, when we said that to get the score of a 4 on our MBO, we had to generate $10 million in new business, that $10 million is a specific number, $10 million. It's also measurable. Presumably we can measure how much new business they bring in. We want it to be attainable and realistic. That is, it's not too much of a stretch goal. Hopefully it won't be impossible for an employee to generate $10 million in your business. And then also you want it to be time-bound. That is, that employee has perhaps one fiscal year, beginning January 1st and ending at the end of that same year, to actually obtain that $10 million in new business, and so it's time-bound. And we don't necessarily think of BARS or MBOs is being one or the other, or all or nothing. Typically we'll have multiple ratings. Some of them will be behaviorally-anchored and some of them will more closely resemble measure by objectives. And so for example, it might try to measure things like customer service or teamwork or leadership or reliability as behaviors that we can observe and that we might tie to the behaviorally-anchored rating scales. And then separately we also might have a separate set of objectives for them to pursue such as new business, business retention, and quality measures that we'll have data for, these output data for. And then, so we'll have a separate section of our performance evaluations for those where the outputs are easily measured even if the inputs that we know correspond to good outputs are not measuring. The second set of best practices I want to discuss are regular communication. And specifically, we want to specify to an employee the desired behaviors and objectives in advance of when they're actually measured. We also want to meet with our employees regularly and to explain scores. Kind of the golden rule of performance evaluations is that they should not be a surprise when the formal performance evaluation comes in. That is, if the manager is communicating expectations and communicating what behaviors or what objectives correspond to what scores, then it's taking some of the subjectivity out of those ratings and therefore making them more expectable and more predictable to employees. And that also helps the manager have more of a learning conversation, a more developmental conversation, while still having these evaluations be used for judgmental purposes as well. So for example, a manager could say something like, I see that you scored a 2 out of 5 on this dimension. Let's work together to see if we can figure out ways to bring that score up in the future. So in summary, tying performance evaluations to pay and to promotions and to terminations and so on, can harm the developmental purpose of those performance evaluations. We also know from a lot of experience that raters will often try to avoid conflict by giving very high evaluations. And we can reduce bias and conflict by trying to benchmark our performance measures and communicating them in advance of the formal evaluation themselves.