Paul Hague
Paul Hague

June 30, 2015

Innovations in predicting behaviour width=

A professor of meteorology jolted his students during a lecture with the following statement: “there is an 80% chance that the weather tomorrow will be like the weather today”. He was making the point that the past behaviour of the wind, the rain and sunshine is a good indication as to what it will be like in the future. So it is with people and markets and we can use this to our advantage.

One obvious way we marketers predict the future is by asking people their intentions. “How likely are you to buy…?”. “Will your purchases go up, go down or stay the same?”. “How will you split your sales between different suppliers in your next purchase?”. However, we are on dangerous ground with hypothetical questions such as these. People can change their mind. Their answer may be given to impress the interviewer, deceive the interviewer or simply because they don’t want to say that they don’t know.

This is not to say that we shouldn’t ask questions about intent; it would be a missed opportunity not to do so. However, there are other ways to predict behaviour and we should consider them. Just as predicting the weather is based on a myriad of evidence, so too we could do the same in marketing.

We can predict four types of behavior with modelling:

  • Price elasticity: alarmingly few marketers know the elasticity of demand for their products and the effect that price increases or decreases have on their sales. One reason for this is that we are negligent in collecting data on the subject. Imagine that every time we had a price increase, we assiduously logged the sales over the forthcoming months, the price changes of the competition in the same period, and the index of inflation. Over time, we would build up a database from which we could model the effect of the price changes determining cause and effect and predicting likely outcomes.

  • New product acceptance: when we launch new products we should record who makes a purchase, at what price, where they are based, what type of company they are and so on.

  • Customer loyalty: we can test loyalty by recording those companies that move to another supplier, the events that led to the move, the company they moved to, the prices they paid, the delivery they received, the number of visits or calls from the company etc.

  • Promotion effectiveness: we can test promotions if we record where and when the promotions take place, how much is spent, what the resulting sales were, what the competitions’ promotions were at the time etc.