What Is the Best Type of Analytics?
Often, when people explain the three types of analytics, they present a diagram that shows descriptive, predictive, and prescriptive analytics building on each other. That is, you need to do descriptive analytics first, then predictive, and finally prescriptive. The diagram then usually suggests that descriptive analytics is the easiest but adds the least amount of business value. On the other end of the spectrum, prescriptive is labeled as the hardest but yielding the most potential business value. Figure 1.2 shows an example of this type of diagram.
Figure 1.2 Sample of a Typical (but Misleading) Diagram of the Types of Analytics
This diagram is visually interesting and can provoke a good discussion. And we, the authors of this book, used such a diagram extensively in the past. However, we now think that this diagram is misleading. In some cases, the diagram can be true. But, it is not universally true. And it may be true in only a small number of cases.
The value of the different types of analytics is tied to the problems you are solving, not the techniques themselves. Each type of analytics can have relatively minor impact on a business or completely change the business.
For example, if your descriptive analytics project is to just come up with a better understanding of what type of product is sold in what geographic area, you will likely get some interesting insight, but it is unlikely to change your business. If, on the other hand, your descriptive analytics project uncovers deeper insight about your customer, it could cause you to change your entire business, as in the case of Circle of Moms. This represents a huge strategic change. As another example, good descriptive analytics applied to data sets concerning cancer patients has uncovered potential life-saving treatments for different types of patients. It would be hard to argue that this type of descriptive analytics isn’t extremely valuable.
Predictive and prescriptive analytics work the same way. You could have many nice projects that are relatively small in terms of the overall strategy of the firm (but could still be good for that segment of the business) and you can have projects of strategic importance. For example, a small predictive project may involve improving forecast accuracy by 10% for certain items. Certainly this is nice, but it’s unlikely to change the business. On the other hand, Amazon’s use of predictive analytics to make buying suggestions changed the nature of retailing.
For prescriptive analytics, we could imagine that the routing of trucks for DC Water saved some money but didn’t change the business. On the other hand, Coca-Cola’s use of prescriptive optimization to blend orange juice to improve the quality of the product could have significant strategic impact on the product. In another example, Indeval, Mexico’s central securities depository, uses prescriptive analytics to help settle securities. The system determines how to match buyers and sellers and clears $250 billion per day, saved $240 million in interest over 18 months, and made the market much more liquid. Indeval says that the system was in place during the stock market crash in 2008, and the extra liquidity in the system allowed people to exchange securities more often during a single day—which was much better than in other countries, like the United States, where the lack of liquidity meant people were stuck holding securities for an extra few hours as the price of the security rapidly decreased. Indeval’s use of prescriptive analytics is obviously very strategic.
It is also impossible to say which type of analytics is easiest to implement. You can have a simple descriptive analytics project where you load the data you have into a better reporting tool and immediately gain insight. Or you could spend two years implementing a full-blown descriptive analytics system that gives you access to every bit of data in your organization. Likewise for predictive and prescriptive analytics: You can do good work in an Excel spreadsheet or you can custom build systems that require years of effort and huge teams of people.
Finally, there is no particular order in which to implement these systems. You might think you need to do descriptive analytics to understand what is going on in the business first. But if you have data (which most firms do) and know what problem you want to address, you can skip descriptive analytics and start directly with a predictive or prescriptive project. And, in practice, this is often what happens. The types of analytics projects you do depend on the business issues you need to address, the value of the project, the ease of doing the project, the skills of your team, and many other factors.
We wish there were a simple roadmap that you could follow to get the most out of analytics. But the wide range of analytics applications and tools is what makes the field so fascinating and rich. You have many different options and have to pick the best approach for your business for each project (big or small).