The only way to know if our IT systems and processes are enabling the business is to measure them. When something in business is measured and made the responsibility of management or a group of people, their behaviour changes.

You may not have thought about it but it’s a fact: Metrics change the behaviour of the people being measured. The trick is you changing the behaviour positively or negatively? The secret between whether you will have a positive or negative impact lies in how well the metric is designed.

So, what is a well designed metric? A well designed metric is one where the objectives of the metric are in line with your business requirements. This means that the behaviour of those being measured tends to be in line with the business requirements. Well designed and measured metrics are method of control.

If a metric is not well designed; not in line with actual business requirements or is not measured correctly, then the aforementioned “control” can drive the behaviour of those being measured in the opposite direction and damage the business.

I’d like to use a non-IT example to illustrate this point:

Non-IT example: The UK Public Health System

In the UK, the National Health Service (NHS) decided that the length of time patients had to wait for operations was a useful metric. To measure this it set targets to reduce the lengthy of waiting lists. The results? Exactly the behaviour the NHS asked for.

The waiting lists got shorter and the targets were met. The only problem was that patients were still waiting for operations. In many cases just as long and in others even longer than before! Hospital administrators were not allowing patients to be entered into the waiting list until their operation had been scheduled to a time within the target waiting time. So a second, informal waiting list was inadvertently created. On it were those waiting to be allowed on the regular waiting list.

Some places just outright lied to people trying to book appointments to make their numbers look better.

This example is a fantastic one for a couple of reasons:

1. It shows how inventive people can be in producing exactly what was asked of them.
2. It clearly illustrates how critical it is to measure what is important.

In this example, it was the waiting list that was seen as the most important thing. The flawed logic of this being that if the lines got shorter, then people must be getting care. There is nothing in this metric that shows a measure of people who are actually getting the care they need. The only thing being measured is that the lines are getting shorter. In this scenario, a better (and harder) measure could have been to measure the time patient’s referral dates against patient operating dates.

The problems of the example given are the same problems we face in IT and in business in general. Somewhere, somehow we’ve lost focus on the overall business objectives. The business objective of the NHS is supposed to be saving lives and helping people. The metrics used were seen as put in place for only the purpose of “catching people out” and therefore, people made sure that they weren’t the ones “caught”.

Your IT metrics work the same. On the surface a metric of “reduction in the number of calls made to the help desk” seems like a good idea but in practice it could be harming the business more than helping it. Let’s take a look at the positive aspects of having an increase in the amount of calls made to the help desk:

• Your help desk is so customer focused that people now call them more than they did before because they know that when they call, they’ll get a definitive answer.
• A new version of software has just been rolled out and people call the helpdesk to get a little extra help on how to use the application because they’ve forgotten a few bits of the training they had on it the week before.
• Your help desk has passed into the realm of an ultimate Service Desk and now they handle more than just breaks and complaints, they also answer business related questions and initiate standard changes such as desktop provisioning and procurement.

We are all happy to work for goals that we think we have a reasonable chance of attaining. If we are asked to do an impossible job like say keep an IT service running 24/7 when we have limited staff, expertise and support, we know that we can only be blamed for not living up to those standards. The natural instinct is to give up and simply go through the motions of trying. Metrics must be set to be seen as achievable and most importantly make business sense to encourage positive change of behaviour.

The moral of this is that metrics must be achievable and recognition and praise must be given to those who achieve and excel in over-achieving against those metrics.

Regards,

Phillip S. Palmer