What Do Roulette and Your Customer Surveys Have in Common?

We all know the gambling game, Roulette. You place a bet and a wheel containing 37 slots is spun and a ball added. The wheel spins and spins until it comes to a stop and the ball falls randomly into one of the slots. But how many times do you have to spin the wheel to obtain an equal probability that the numbers 0 – 36 appear? The answer is around 1500 spins (calculated using fancy statistics)! Although this problem is based around Roulette, we can apply the same principles to business.

For example, how many surveys will your company have to send out to get a good representation from each of your customer target demographics? Or to put it in a more scientific way, how many surveys will you have to send out to get an equal representation of returned surveys from all demographics.

We can break this problem down into its component parts. Let’s say your company appeals to two age demographics: 18-30 and 30-40 years old. In addition to this, both men and women use your products and you have customers in 20 London postcodes. That means you have:

2 x Age demographics

2 x Gender demographics

20 x Postcodes demographics

Totalling 80 unique demographics!

So if you assume that each demographic is equally as likely to return these surveys, then you would have to send out over 5000 surveys to get a good return! By good return, I, of course, mean an equal return rate across all demographics.

But why is it so important to obtain an equal representation from each demographic? If you were to ask one person what their favourite film is, would that be a fair reflection on what everybody’s favourite film is? Of course not. If you were to ask, say 1000 people what their favourite film is, this may give you a better idea of what the most popular films are. With the example above, one male 25-year-old customer from NW4 may not reflect the views of every male under 40 living in NW4.

Using scientific principles to design business strategies will allow you to make predictions on the outcome of these strategies. Such thinking will help you make better business decisions, as in the case above, where you will know prior to sending out surveys, how many you will need to send out to get a good representation from each demographic.