Who Are Your Gold Medal Customers?

How does predictive modeling work? A recent article describing how to predict medal counts by country yields some interesting parallels with our own best customer predictive modeling process.

Researchers at a German university recently built a predictive model to determine which countries will win the most medals during the 2012 Summer Olympics in London. The model they built took into account several factors, including economic, demographic and cultural factors. Their primary finding, which is not surprising, is:

Money plays the biggest factor in which countries win the most medals

However, notably, the researchers found that a host of other factors also play into a country’s final medal count at any given Olympics, including:

1. Population: larger countries have a larger pool of athletes from which to draw

2. Political system: socialist countries put more national resources into supporting athletes

3. Home advantage: athletes competing in their home country tend to win more medals

4. Climate: athletes from countries with more extreme climates – hot or cold – are at a disadvantage relative to those from countries with moderate climates

Another notable research finding: certain countries, like Jamaica and Kenya, are able to buck the trend. In statistics-speak: these countries are considered outliers, meaning that the factors related to these countries (i.e., the variables) indicate that they should be winning far fewer medals than they actually do.

The story of this German university’s research into the factors that predict Olympic medal counts by country parallels very closely to that of the numerous “best customer modeling” projects in which MindEcology has engaged over the years. Translating the Olympic medal story into the world of predictive model-building for finding best customers, here are some key parallels, point-by-point:

1. Not all countries have the same chances of winning a medal at the Olympics.

Parallel: Not all of your prospects are equally likely to buy from you.

2. There are identifiable factors that can predict a country’s medal success.

Parallel: There are identifiable factors that can predict who will be most likely to buy from you.

3. Money matters, but it’s not the only factor in winning medals. (Case in point: Japan has the world’s third-strongest economy, but it is predicted to come in 13th in terms of medal count).

Parallel: People with more money have greater capacity to buy from you, but having money doesn’t necessarily mean they are the ones who will buy from you. Several other factors matter, and sometimes capacity to buy might be inversely correlated with their propensity to buy from you.

4. Good models can be highly predictive of Olympic success

Parallel: Good customer models can be highly predictive of customer response rates.

5. There are always going to be outliers that buck the trend (i.e., Jamaica and Kenya), but these don’t take away from the power of the model.

Parallel: These models work; they can save you significant amounts of money by helping you focus your marketing efforts on those prospects likely to convert. Sure, there will be customers who buy from you who don’t match the model’s best customer characteristics, but that doesn’t take away from the power of the overall model. Think “law of large numbers.”

The bottom line is: there are ways to identify your own gold medal customers. Take advantage of predictive modeling before your competition catches on.