machine learning vs. rules

Whether you’re just getting started with personalization or you’re an experienced pro, you’re probably familiar with the two key ways to implement personalization across your digital channels: rules and machine-learning algorithms.

One approach is not inherently better than the other. Rather, they each have their strengths, weaknesses and ideal use cases. A strong personalization program makes the best use of each of these approaches. Let’s describe them each in a little more detail.

A Real World Example

To illustrate, we can draw a parallel to the real world. Let’s say you have a receptionist and there’s some important information you want him to convey to employees when they enter the office. Specifically, you have a few messages for him to share with some individuals today:

  • “When the CFO comes in, tell her she needs to dial in to this meeting as soon as possible, and offer to bring her some coffee.”
  • “When John comes in, give him his wallet. It turned up in the lost-and-found.”
  • “When anyone from engineering comes in, remind them that we had to move their morning meeting to a different conference room.”

The messages described above are basically rules. They are manual (because you have to tell the receptionist exactly what to say and whom to say it to), but they are extremely valuable when you have something very specific you need to say to certain people.

Now, let’s say, in addition to these specific messages for him to relay, you give him some guidelines for how he should generally greet employees as they come in to the office. You encourage him learn about each individual and decide for himself what to say to each person. For example, you could suggest that he get to know each person’s interests and communication style and respond appropriately to each person. And as he gets to know people over time, the relevance of his interactions will keep improving.

Machine-learning algorithms are like these general guidelines you’ve given the receptionist of how to interact with the employees. You set up the algorithm, but you don’t tell it exactly what to do. It uses the data it has available to deliver the most relevant experience.

Both rules and algorithms play extremely valuable roles in your personalization strategy. The trick is to know when to use one over the other – and when to use them both – to help you develop the best experience possible for each of your visitors.

A Comprehensive Guide to Machine Learning and Rules

We want to provide you with a complete understanding of rule-based and machine-learning personalization so you can successfully apply both in your personalization efforts. Download our latest eBook, The Marketer’s Guide to Machine-Learning vs. Rule-based Personalization, to receive a comprehensive course in both forms of personalization, including definitions, examples, ideal use cases, strengths and weaknesses, and tips on getting started.