Let the Data Change Your Views: The Bayes Method

4.0. Introduction: What Is the Issue?

Most business outcomes are unpredictable and yet we often have to make a decision that depends on that outcome. To do so systematically and objectively, we need to describe the uncertainty of the outcome with probabilities. For instance, we may judge that a new product has a 50% chance of succeeding. But this 50% is not set in stone. It may change if we obtain further information. For instance, if the product receives a positive market test, or if the main competing product is recalled or if the economy hits a sudden boom, then the probability of success should surely be revised to a value higher than 50%.

There is a systematic method for updating a probability to reflect new data of this kind. It is called the Bayes method and is very easy to implement. To apply it, you need some extra probabilities relating the data to the business outcome. These may not always be easily available but if you are going to do anything better than guess the chance of business success, you will need to obtain or estimate these probabilities.


Titan Resources is searching for sites with minable levels of titanium. From rough geological and satellite data they can narrow the field of likely sites greatly but even among these more promising sites it is estimated that no more than 1 in 500 will be minable. To test a site, a core sample is taken and analyzed. It is ...

Get Data-Driven Business Decisions now with the O’Reilly learning platform.

O’Reilly members experience live online training, plus books, videos, and digital content from nearly 200 publishers.