But there is more to it than that. Here is my approach to speed.

You must be able to see which elements are standard and which require special attention.

Some elements of a job are standard. Once the setting out has been decided, there may be hundreds of operations which present no new problem. They have been solved already, either on this job or many times before.

Others contain some uncertainty, establish something that follows, or have consequences which make an error expensive. These two kinds of work should not be done at the same speed.

The obvious mistake is to treat something unusual as routine.

The less obvious one, and the one I am probably more prone to, is to treat routine work as though it were unusual. It is tempting to call this carefulness. But if an operation is understood, repeatable and adequately controlled, giving it additional attention does not necessarily improve the result.

Slow down for the special. Learn to leave the standard alone.

Recognising something as standard is of little use if every time you encounter it you do it differently.

A standard operation needs a standard procedure: the same sequence, the same tools, the same checks, preferably in the same order. There is usually more than one correct way of doing something. But once I have found a way that works reliably, I should have a good reason for departing from it.

Otherwise I keep solving problems that I have already solved.

Much of the time lost is not spent working slowly. It is spent doing things around the work.

Before beginning, I try to ask a simple question: do I have everything I need to finish? This includes decisions.

You start an operation and discover that something is missing. You need another tool. A cut has not been made. The tiles are in the wrong place. You reach a detail which should have been resolved before the adhesive was mixed.

Every interruption creates movement: getting up, walking somewhere, finding something, putting one thing down and picking another up, then returning and starting again.

If I stop tiling to make a cut, I put down my tools, clean my hands, measure the tile, walk to the saw, make the cut, come back and begin tiling again.

The cut itself may take less than a minute. Do this once and it doesn't matter. Do it twenty times and it does.

Where possible, measure several cuts together. Make several cuts together. Prepare the tiles before mixing adhesive.

There is no point making ten cuts in advance if the later dimensions depend on tiles which have not yet been fixed. Efficiency must not get ahead of certainty.

But where operations are genuinely independent, group them.

“Measure twice, cut once” is good advice for something unusual, difficult or expensive to get wrong. It should not be the normal way of working.

For ordinary measurements, I try to discipline myself to give the operation my full attention the first time. Stop whatever else I am thinking about. Measure. Read the number properly. Remember it or write it down. Then move on.

The important thing is not to measure twice more quickly. It is to make the second measurement unnecessary.

A driver cannot normally make a decision carelessly and then go back and check it. The attention has to be there when the decision is made.

Repeated checking feels safe, but it can become a substitute for concentration. There will always be things worth checking twice. That takes us back to the first rule.

At the end of a slow day, it is easy simply to conclude that I was slow. That tells me nothing.

Where did the time go? Was the work genuinely difficult? Did I treat standard elements as special? Did I keep changing operations? Was I badly prepared? Was I walking backwards and forwards? Did I check things which did not need checking? Did I spend two hours correcting a mistake?

Sometimes the work simply took as long as it needed to take. Perhaps the mistake was in my expectation.

In the middle of a job, every individual action can seem reasonable. It is only afterwards that a pattern becomes visible. Otherwise experience is just repetition.