02.02 – Measuring Results and Behaviors

Recall that in course one, we talked about setting smart goals and objectives. These goals and objectives have to be linked to the strategy and vision of the organization. That’s an absolute first key step. It’s very important. Then we need to measure people’s performance and behaviors against those objectives. Now some of these objectives will lend themselves to quantitative measurement, others will not. For example, it might be easier for me to quantitatively evaluate your sales numbers or your cost-cutting numbers. But in addition to hitting those numbers, companies also care about you exhibiting certain behaviors.

Such as you hit those numbers in a way that supports and reinforces corporate culture, that you grow yourself as a leader and you develop others along the way. And ultimately, these dimensions of performance, competencies, and behaviors feed into your performance appraisal form. This is a segment of a typical performance appraisal form. In most companies, managers would fill them out once a year at the end of the year. Many companies are now transitioning toward feedback systems, where such information would be communicated to employees more frequently than once a year. Accenture, Microsoft, Amazon, Netflix are good examples. But let’s go back to this performance appraisal form for a minute.

You can see that this employee is being evaluated on five different dimensions, project management, customer service, production support, teamwork, and communication skills. So the first step is to assign a quantitative score to the employee on each of these dimensions by comparing his or her performance and behaviors relative to standards and objectives. So for example, for assigning a score for project management, your manager might consider goals of completing projects on time and on budget. For customer service your manager might consider scores from customer satisfaction surveys. Lexus, an automotive manufacturer weighs customer satisfaction scores very heavily in its appraisals of dealers. For teamwork many companies would go out to your peers to get that evaluative feedback.

Some companies would rely on input from your primary manager.

In addition to assigning scores on each of these dimensions, your manager would also need to provide some vignettes or examples of behaviors and outcomes to explain those ratings. And even if your company does performance appraisals once a year, don’t wait until the end of the year to come up with a list of those behaviors and outcomes. In the beginning of the year, start a text file on all your direct reports, your teammates you need to evaluate. And write down their key behaviors and outcomes following the completion of major tasks and projects.

In addition to getting scores on Individual dimensions here, an employee would also receive an overall performance score. Now in some companies it would be either a simple average or weighted average of the scores on different dimensions. If it’s a weighted average, it would typically be weighted by the importance of different objectives or different dimensions. In other companies you would rely a bit more on managerial discretion, meaning that the manager would come up with a holistic score that they believe best describes your overall performance, of course in light of the ratings on the individual dimensions. This overall score is really important. It determines your likelihood of promotions, your salary raises, your compensation, your bonuses, and so on.

A quick sidenote on the scales, most companies use scales that vary from three point to seven point scales to measure behaviors and outcomes. It’s fair to say that on most behaviors and outcomes, a five-point scale would be sufficient to capture variations in behaviors and outcomes. It is less complex and more easily interpretable than a seven-point scale. There’s also research showing that five-point scales can be more motivating than the three-point scales. Because employees are more likely to believe that they can move up on a five-point scale than on a three-point scale.

So at this point, I want us to pause and recognize that there are two fundamentally different approaches to assigning ratings to your employees. One is what we call an absolute rating system. What I mean by this is that when I evaluate a given employee, I don’t compare that employee to other employees in the organization. I evaluate them independently. And the second approach is a comparative rating system, where I evaluate every employee by comparing him or her to their peers.

The most typical form of a comparative rating system is a force distribution, or force performance curve. Recall that in school or college you had some classes that were graded on a curve where say, no more than 15% of students could get an excellent score. A forced distribution approach is very much similar, where managers assign ratings to employees in accordance with pre-specified performance distribution percentages. So the classic example here is GE’s forced distribution, which was used in the 80s and the 90s when Jack Welch ran the company. This curve required that no more than 20% of managers were assigned to exceeding expectations, the top bracket.

70% would be assigned to the middle bracket, meeting expectations, and 10% had to be assigned to the lowest bracket, not meeting expectations. Those would be employees that would be the first to leave the company if their performance does not improve. The top 20% would be in line for the biggest salaries, the biggest bonuses. Jack Welch referred to it as the vitality curve. And he called it the kindest form of management which is a very interesting statement when you think about this. Now Jack Welch’s logic was that very often we’re fired from companies for inexplicable reasons, and this approach gives you a lot of useful information on your performance.

It gives you an opportunity to change your behavior, and you get a head start on your job search if you don’t improve. So, you call that the kindest form of management because you get a lot of useful information in your performance. You know exactly where you stand in your organization. You know your prospects with a company, and you get a lot of that information well in advance of major decisions with respect to your career. This approach to evaluating performance, forced distributions was really popular in the 80s, 90s, and early 2000s when lots of companies rushed to emulate GE’s approach. The enthusiasm for forced distributions has subsided significantly in recent years and for good reasons, as we’ll talk about shortly.

But there’s still a lot of companies that are avid advocates of forced distributions. Lending Tree is a good example. It’s a leading online mortgage provider. Their curve is actually a little harsher at the top, where no more than 15% of employees can be assigned the top rating, which is a one. 75% should be assigned a rating of two, meeting expectations, and bottom 10% a rating of three, not meeting expectations. Ones receive the largest bonuses. Threes are the first to be fired if their performance does not improve. And so, in performance appraisals at LendingTree, the key question you’re wondering about going into your performance appraisal is, are you a one, are you a two, or a three?

GlaxoSmithKline, a leading pharmaceutical company, Goldman Sachs, a premier investment bank, all of them use variants of forced distributions. This is another example of a forced distribution used by AIG, a leading insurance company. Top 10% at AIG get a rating of 1, the top rating. The next 20% are 2s. The next 50% are 3s, and the bottom 20% are 4s and 5s. As you might have guessed, the 1s get the largest bonuses, the opportunities for promotion, the largest salary increases. Bottom 20% might not get any bonuses at all, and the 5s are the first to be let go if their performance does not improve.

Another variant of a comparative rating system is a forced peer ranking, where a manager rank orders his or her teammates, subordinates, direct reports, from the best to the worst.

Some consulting companies still use that system where at the end of each project, manager would rank order all of the consultants from best to worst. And your end of the year score would be essentially a weighted average of these scores where the weights are the times you spent on different projects.

A variant of a forced peer ranking or even a way to get to this ranking is a system of pairwise comparisons. Where instead of assigning one overall ranking right away, I conduct a system of pairwise comparisons, where I compare each teammate’s performance to the performance of every other teammate. So here, I would take Franz’s performance and compare his performance to Jenna’s, Aisha’s, Michael’s, Luciana’s, and Qinlin’s. Then I take Jenna’s performance, I had already compared her performance to Franz’s. So I compare Jenna’s performance to Aisha’s, Michael’s, Luciana’s, and Qinlin’s. In this particular case, you can see that as a manager, I will have to make 15 pairwise comparisons.

The formula is n multiplied by n minus one divided by two, where n is the number of teammates. So in this particular case, it’s 6 multiplied by 5 divided by 2. And then I simply count the number of wins for each employee. The person with the largest number of wins is propelled to the top of the ranking. The person with the fewest number of wins is at the bottom. And you have a nice tie breaker, which is a head to head comparison in case you have a tie.