Mad Management: Managers must use data, not incentives

editorial image

The single biggest management fallacy is that your job is to manage people. You don’t need to manage people. I repeat: You don’t need to manage people. By the time they come to work they are pretty grown up, so why are they so often overseen as though they are kids.

Today, when we have the most educated workforce ever and so many people desperate to work, and want to do a good job, we have some of the most over-managed, top down organisations ever; especially in the public sector.

So, what is the manager’s real job? It is to design and manage their organisation so that people can do their best work. It is not about governance. (Will someone please tell HR.)

That is why, when Dr Deming returned from his great success in helping bring about Japan’s manufacturing revolution, he declared that the West needed a leadership transformation, because they had no idea how good organisations really worked. In fact he said “new knowledge” was needed. When the USA companies finally realised that the Japanese were eating their lunch because their quality was so much better, the panicking executives rushed to ask Deming for the remedy. Deming said: “If I could reduce my message to management to just a few words, I’d say it all has to do with reducing variation.”

What ?! Not incentivising the workforce, or IT solutions or costs-cutting, the panaceas peddled by the big consultancies? What is this “variation of” anyway? Well, work takes place in a system comprising a series of interconnected processes that end up delivering a product or service. These processes fluctuate naturally, be they cold-rolling steel, serving a customer, or travelling on a bus. Commonsense tells us that no steel rod from the same strip is identical, neither are two hamburgers or the time of bus rides from Hunters Bar to the city centre nor the ambulances to an emergency. In other words processes vary, and so, therefore, do the outcomes.

The way a process performs should be determined by the need. Medical instrument manufacturing demands much tighter tolerances than does wheelbarrow production.

The process must have much less variation because the upper and lower specification limits need to be very narrow, measured in microns, not millimetres. Japanese trains arrive and depart to the second, but as a bus passenger I don’t really mind whether it takes 8 minutes or 12 to get to the Crucible; but this would be unacceptable in an ambulance emergency call.

Why reduce variation? In the first place, because it affords better prediction. For everyday activity people want reliability, consistency. Take this ridiculous waste of money on HS2 to gain 20 minutes on a trip to London. Is that what passengers were complaining about? No,they wanted to have confidence that train that would reliably get them to their destination on time so that could plan their day. But no, George Osborne had to have his “project” with which to make his mark. Billions spent on the wrong thing, again!

How do we get this reliability? By improving the process firstly through reducing the variation, the first step of which is to establish its capability against the real demand. Let’s say the number 88 bus passengers have now demanded a faster and more reliable journey, i.e. to get to the Crucible in between 6 and 8 minutes. 
No amount of exhortation to the staff is going to achieve this. All that will happen is that cynicism about management will increase exponentially.

The system has to improve. What the company would need to do is to measure the current performance over time of the 88 bus. The method for doing this is a capability chart, more commonly known as a Control chart, which gathers all the information on the journey times, say over a year, and works out statistically if it is capable of delivering that goal reliably, i.e. under 9 minutes with an expected failure rate of not more than 3 per 1,000 journeys. Not surprisingly the chart will probably show that the system is not capable. In fact, you could expect a normal variation of between 5 and 20 minutes, even though most the time it could get you there in under 8 minutes. So, if management announces that it will meet the demand but does not address the system, there will be a lot of unhappy passengers and staff, because it simply will not work.

As Deming would say, “The process has spoken”.

This is exactly what the government did when it announced the 4 hour target for emergency treatment. Typically many hospitals “failed” and, of course, the workers were blamed by both the politicians and patients.

(The picture of control chart above shows how setting a target without reference to the capability of the system ensures target “failure”. In this case at a rate of about 90%. Morale collapses.)

What must management do? First study the process, and the only people who can help them are the workers in the process. As the Japanese would say, on the shopfloor is the truth.

Causes of bus delays often can be explained by the drivers and controllers, e.g. school runs and increased use of pedestrian crossings. Significant trends may also be detected.

A very important step has been taken. Management now knows the reality, thanks to the control chart and staff, and can begin the journey to really improving quality and costs.

Next week I will outline how this is best done.