Monday 21 September 2015

The art and science of marketing forecasts

Every marketing plan requires some kind of marketing forecast--usually, sales forecasts on a product by product OR brand by brand OR market by market basis. But how do you develop a reasonable forecast, especially when you're looking at something new or somewhat unpredictable?

Tetlock and Gardner, authors of Superforecasting: The Art and Science of Prediction, researched people who are more often correct than incorrect in their forecasts, and concluded that three things can improve a forecast:
  1. Don't make up your mind in advance. Instead, be open to new ideas and critically evaluate information. If you want to examine the market for doughnut shops in London, media coverage of National Doughnut Week's top trends can help you think about the ascendancy of doughnuts after years of the cupcake craze, for instance. Try not to rule out information that contradicts your initial thoughts on the forecast--instead, ask hard questions to test the data before relying on it.
  2. Identify the various elements that can influence the forecast. It's difficult to predict exactly how many doughnut shops will open in London next year, but easier to start by calculating the number of residents of London, the number who typically buy doughnuts and the number of doughnut shops already in existence. These statistics will help you consider what drives trends affecting consumption and business formation.
  3. Look at what others are forecasting. Find authoritative sources of information, examine their forecasts (element by element if possible) and see how your forecast compares. Here, you might track the number of new London-area shops opened by Krispy Kreme and other doughnut retailers, analyse how the number of new openings has changed by chain over the years and link this information to other developments in the marketing environment that affect doughnut consumption and retail expansion.