The perverse effects of KPIs


Guest writer, Anneleen Vanlommel,
Employee Engagement Enthusiast with Herculian Alliance.


Anneleen is our former colleague and now she combines employee engagement with customer experience.

Not everything that can be counted counts, and not everything that counts can be counted. 

– William Bruce Cameron

I weigh 62 kilograms. Five kilograms too much to my liking. Time for action! To know if I’m actually getting results, I weigh myself every Monday morning, right after showering and completely naked, to make sure that I can work with comparable results. I’ve been doing that for seven weeks in a row. Without any result. What am I doing wrong? I explain that to you in this blog because we all make the same mistakes in defining our KPIs.

Measuring what doesn’t matter

You use Key Performance Indicators (KPIs) to analyse performance. Thanks to KPIs you want to induce change. And they are perfect for that, because a KPI impacts the behaviour of your employees. Someone measures it, so it’s important. You don’t even have to link an evaluation or remuneration to it. A KPI changes the behaviour of your employees. Point.

And that knife cuts both ways. Because when you define the wrong KPI, you get the wrong behaviour. And we still see that happening too often. Employees in a call centre know them well: the KPIs with perverse side effects. Managers in a call centre might evaluate their employees on the length of the phone calls they are doing. In an ideal world they would take less than 3 minutes, because then they can help more customers, keep the waiting time under control and have more satisfied customers. The annoying side effect? Employees who, after 2 minutes and 50 seconds, tell the customer they’ll call back in a minute. The KPIs are green. The customer turns red.

Measuring what matters and still overshooting the mark

Soon I found out that I’d rather weigh myself in the morning than in the evening, that I’d get a better result if I didn’t drink any water the night before, and that it was to my advantage to go to the toilet first. Once I saw the number 59 appear. And quickly disappear again. A typical phenomenon: measuring the intended result but not actually achieving what you want.

Your goal is customer satisfaction. So you show your employees on a weekly basis how often they make a customer happy. You measure what matters, and yet you create problems. A small selection of what we see happening: 

  • Employees who only send surveys to customers they expect to be extremely satisfied.
  • Assigning colour codes to the survey so that customers are well aware that a 7/10 or 8/10 answer is only ‘orange’.

  • An employee who asks your customer to give a high score, because he will be evaluated on it.
  • Or handing out gifts to customers, hoping for better scores.

Measure behaviour, not results

Most managers focus on results. Sales managers focus on sales, service managers focus on customer satisfaction, parents focus on report cards, people who diet focus on weight. But by measuring results, you won’t get there. In order to motivate your team and achieve success, it is better to do a behavioural measurement. Here is the difference:

A result measurement

  • tells you if you have achieved the goal
  • can’t be directly influenced by you
  • gives you information at a moment you can no longer change the result

A behavioural measurement

  • tells you how likely it is that you will achieve your goal
  • can be directly influenced by you
  • gives you information when you can still influence the result

Therefore, a behavioural measurement has a more motivating effect than a result measurement.

Is your goal customer satisfaction? Then your result measurement shows how many satisfied customers you have. Your behavioural measurement, on the other hand, measures what your team does to ensure that customers are satisfied.

Why is it that we so often use the wrong KPIs?

If it’s so simple, why do we always use result measurements? There is one simple reason for that: they are easy to implement. You don’t have to look for them, they are often delivered to you automatically.

Measuring behaviour is a lot more difficult. And there are two reasons for that:

The behaviour with a direct impact on the result is difficult to determine.

How do you determine which behaviour has an impact on your intended result? Many managers do this intuitively. They start from assumptions and select those elements that they think have a direct impact on the goal. Like answering customers quickly, for example. If we do that, customer satisfaction will increase. And so we measure the speed with which we solve customer questions. Employees adjust their behaviour and focus on the speed, not on what they should do to make the customer satisfied. A world of difference, because satisfaction is rarely about how fast you do something for a customer. Your customer just needs to know who is going to do what and when. No vague promises. Customers like it super concrete.

Behaviour is difficult to measure.

Have you figured out what your employees have to do to create satisfied customers? Then it’s a matter of measuring how often they behave in this way. How often did your employee give the customer insight into when his question will be answered? Quite a challenge, because this data is seldom available and collecting it takes some effort.

How do you determine a good KPI?

Find the desired behaviour

First, determine your goal. What should your customer say about your organisation? Then, experiment to find out what the behaviour is that your employees have to display in order to reach that goal. Are customers more satisfied when we call them, rather than emailing them? Do customers want a fast solution? Or do they prefer to be kept informed while we take some time? You only know the answer when you test it on a number of customers.

Most important criteria? Your employees must be able to have a direct impact on the desired behaviour. Whether they do it or not, depends solely on their own will to do so. 

Measure the desired behaviour

Is it clear which behaviour is needed to achieve the desired result? Then you have to measure it. And you have to make sure that your employees have an overview of the status. Not a monthly report with the results, but a direct view on how your employees behave.

Do I want to lose those kilos? Then I’d better stop measuring just the result. I have to measure my behaviour. And that’s why I first determine which behaviour has a direct effect on the kilos that appear on my scale. In order to lose weight, I can take two actions: eat healthy and exercise. 

And I can measure those two. How many calories do I eat every day and how many calories do I burn? Do I keep a good record of these two things? That way I focus on what really matters. Not an easy exercise, but better no KPI than a bad one. 

By now I weigh 57 kg. I practice what I preach.

Intrigued by this subject? Then be sure to read The 4 Disciplines of Execution van Chris McChesney, Sean Covey en Jim Huling