Friday, 4 February 2011

If you can’t measure it, you can’t improve it

Does a ruler set your heart racing? Would a new protractor make you feel a little giddy? Did you go into maths lessons with a hop, skip and jump when you were at school? No? Well me neither, but that doesn’t mean to say that measurement isn’t one of the most important aspects of the highly sophisticated lives we live.

Lord Kelvin pointed out that “If you can’t measure it, you can’t improve it”. Those carefully chosen words seem self-evident to me, yet measurement is not something we accept naturally. Much human communication is though words, expression, tone or voice, touch and emotion – all of which are difficult to measure.

But all great advances in human endeavours have come about through our ability to measure things. The list is very long: healthcare, safety, architecture, transport, energy efficiency, food quality, and much more.

A couple of weeks ago I went to see what might be the world’s oldest clock, which could date from the 14th century. It currently stands in the magnificent Salisbury Cathedral which is well worth a visit even if you have no interest in early time pieces. Of course in the 14th century, and many years afterwards, accurate timekeeping was not possible. If you don’t know what time it is, you cannot co-ordinate people or events around you or across town. Nor do you know how long things take to do (like cook a soft-boiled egg or travel to Leeds).

Whilst it is interesting to look back at advances in measurement and understand the benefits they have brought, it’s also interesting to think about the things we can’t or don’t measure. Of course that list is a great deal longer than the many things we already measure. From my limited experiments with measuring time spent on different activities (focused work, day dreaming, wandering around) or the number of data-driven decisions we make in a day (not many), I concluded that it is possible to usefully measure more than we currently do, even though it is difficult and somewhat unnatural.

Benjamin Franklin famously tried to measure thirteen things that he considered important: silence, order, resolution, frugality, industry, sincerity, etc. He tracked his adherence to these virtues at least until he was 79 when he wrote about them, and resolved to follow them longer as they had brought him so much happiness. None of them are concepts that lend themselves easily to measurement. Yet if Franklin hadn’t measured his adherence to them, he would probably have forgotten them long before his 21st birthday.

Bernard Shaw once said that all progress depends on the unreasonable man. I suggest he was wrong; progress belongs to the man with a mind to measure.


  1. I agree with the basic thesis of "if you can’t measure it, you can’t improve it". But I disagree with your conclusion about progress. There's a difference, IMHO, between improvement (which tends to be incremental) and progress (which is often discontinuous or disruptive). Sometimes, progress takes place by replacing what was there by something entirely new; and the measurement applied to the old may no longer be relevant (e.g. no point in making a better slide rule when the calculator has just come on the scene). And to force that kind of change requires an unreasonable man.

  2. I would argue that if you can't evaluate it, you can't improve it. Franklin may have measured silence in the sense of how much of it there was. But what was the quality of the silence? Was it punctuated by birdsong? Was it heightened by the smell of flowers? Measurement is crude. So 75% of an organisation's staff say they support an organisation's values. Tick. 100% of check out staff in a supermarket smile at customers. Tick. But the only way of finding out about what people's values actually are, or how they relate to the organisation's values, or why they behave the way they do is to ask them. That's not measurement. It's qualitative evaluation. Anyone can measure. Only those who are fearless in their quest to get to the bottom of reality can evaluate.


  3. Measurement to me is the ability to build models from the data so that it becomes a predictive tool. The model can then be tested by further measurement if necessary. If we understand the model we can then design improvements, and test those by further measurement. Is this not the basis of the scientific process.

    Measurement is also used in control where we are using a model to provide control actions. The model may be very simple eg a thermostat. Again understanding the thermostat we can think about how to improve it. PID control strategies for example.