How Averages Can Be Dangerous

How Averages Can Mislead

We use averages in our business life all the time, and they are usually pretty straightforward. However, when it comes to making business decisions involving customer data, averages can be misleading.

Let’s say that we had a sample of 100 gamblers from your customer re wards system. From your current database analysis tool you can learn that their average income is $55K per year. So from that information, you might assume your typical customer has an income of $55K and act on that assumption, by building a marketing campaign targeting customers with an income of  $55K.  When you imagine averages, you probably think about a curve like this:

 

 

The problem is there are other kinds of curves with the same average as the one above, but they have wildly different distributions and implications for your marketing campaign. For example this chart shows a flat or what statisticians call a uniform distribution which would be bad news for your marketing campaign.

 

If you are an optimist about your gambling population, you might think you have a narrowly grouped base around $55K like this chart.

And if you are unlucky, some day you might experience this unfortunate distribution and none of them have an income of $55K as seen in this troth shaped bi-modal chart.

 

The important thing to note here is that all these graphs have the same averages, so if you were relying on averages alone to make decisions for your business, your success would be at risk. With misleading data, you may be spending thousands of dollars marketing to unprofitable customers and overlooking potentially valuable ones.

Distributions Matter

Averages don’t matter as much as the distribution around the averages, when it comes to finding ideal customers. Distributions give you a fuller picture of where your customers are and just as importantly― where they are not.

The way to the right people is to use more information, specifically, to use all of the distribution and not just the average. Predictive analytics is the clear winner for making more profitable decisions, because it uses all of the information in the distribution. Predictive analytics is a more complex way of looking at potential consumers, taking into account their past behavior and predicting whether or not they would be receptive to the product or service you have to offer.

To use predictive analytics, you could start studying distributions, send your staff to statistics classes or start benefiting immediately from Stics. The professionals at Stics have made it easy for people who aren’t knowledgeable about statistics to get the important information they need for their company. Stics has predictive models that can give you valuable information about your customer information― which will lead to more profitable marketing campaigns.