This month, M Moser’s Steve Gale makes the case for more maths.
The creative industry is light on people good with numbers. The two don’t seem to go together – you go one way or the other. Why be an accountant when you could be having fun designing stuff? There is truth in designers not liking numbers.
Does this matter? Who cares if your architect can’t do long division? We are paid for ideas, images, synthesis and innovation. To hell with your engineering calculations – computers do that, right?
Wrong. But that’s not really the point. The understanding of numbers and what they say is central to every creative activity, it’s just that we can usually separate it into a separate discipline for others to do. The division of labour enables us to play to our strengths, so volumes, rates of change, distribution curves and binomials go into the engineering box.
I have recently been forced to face some demons in my number-light world, and the demons can be found in every brief that is written.
In any commercial building we calculate how many desks the business needs, the size and number of meeting rooms, the number of people actively using a building during the day, and the demand for simple hospitality, like coffee in the morning. This seems like simply adding up, it’s not real maths, it’s arithmetic – sums. God, I wish it were that simple.
The numbers needed to define a simple office scheme are riddled with problems, but they rarely get interrogated by our dear clients, or worse, by us, the number generators. Here’s some examples described in three orders of complexity.
“It’s really difficult, so we guess, and apply specious rules of thumb handed down in design standards with biblical reverence.”
First order is the deceptively straightforward issue of headcount. As a cornerstone of any commercial brief it is often hopelessly unreliable. It wobbles around the numbers of employees, contractors and visitors, modified by secondments, travellers, absentees and job shares, and further fogged by its place in time, which oscillates during the day, and needs to be placed in the unpredictable future – and quite a distant future, not just day one of occupation. This makes for big error bars – fuzziness, in other words.
The next order of complexity is the way we estimate things like the number of desks needed for this inaccurate number of people, especially a group of agile workers who are not around all the time. When we agree that a desk for every person is a waste of money and space, what is the right or optimum number?
This is a simple statistical challenge, but do we draw a distribution curve and calculate the standard deviation and report the result that satisfies 95% of events? Do we report degrees of confidence in our numbers? Obviously, we only do this very, very rarely, or in my experience, never.
Finally, a really difficult area is estimating things like meeting rooms for a business. These expensive, space-consuming, light-blocking blobs are vital, but the way we calculate their number and size is opaque. I have never heard a client say they have got it right, so it’s very fertile ground for improvement – but it’s really difficult, so we guess, and apply specious rules of thumb handed down in design standards with biblical reverence.
Well worn paths exist to resolve these problems. Airlines predict flight frequencies and plane sizes, hotels calculate room numbers and prices, and they do not ever guess. They use mathematics and statistics for modelling, tweaking and testing over time, getting better each day, to stay in business. Where is this rigour in our briefing routines? Is our guesswork so refined that it’s now ‘good enough’? I do not believe it is anywhere near good enough.
This is not original thinking. Other industries build algorithms to test and improve their output, on a daily basis. Basic design needs a numerate heart transplant, or more people with a numerate heart to help make these core decisions. Find these people and keep them close.