Skip to content

By using or further navigating this website, you agree to Sitecore’s use of cookies. Click here to see our cookie policy.

Sitecore Sessions

How your digital tests benefit from Sitecore’s machine learning

Machine learning helps discover key audience segments

By Lars Birkholm Petersen, Sitecore VP of Business Optimization
Video transcript

Hi, and welcome to another Sitecore Session. My name is Lars Petersen, VP of Business Optimization at Sitecore. If you have been watching the Sitecore Sessions, you know we have been talking a lot about optimization, and very specifically about testing and personalization. Where do you start and where should you start?

Before we dive into that, let’s take a look at how Sitecore works with machine learning for optimization to essentially inform you about where you should personalize, and for which audiences. Let’s say you have just concluded a test; based on concluding a test, you know which variant of that test is most effective in terms of business value of your organization.

Understanding your test results

Here are all the different visitors that have been exposed for the test. Now the machine learning in the background will group those different visitors based on Engagement Value, based on the outcome you are optimizing for, and those different audiences that are detected here. In this case, the machine learning in the background will group them into three different groups, three new audience segment, and that will bring intelligence about some of the audiences that actually prefer other variants than the test winner. This helps you inform the personalization, and it helps you optimize even further.

So let’s take a look at how that works in Sitecore. In this case, you have concluded the test. Now the machine learning comes up and actually has identified two new audience segments: visitors from Germany and visitors who are male. Of course, this is based on the test on this specific site. All we need to do now is use this and turn it into personalization. Then we get much more value based on the Engagement Value that is important to achieve our strategic objectives.

Lead with personalization

So are there any cases why you shouldn’t lead with a test and you should start directly with personalization? Yes, there are. Those would be examples of audiences where you know that they have very specific intent, very specific needs by coming to your digital channels. Now starting with personalization for them, according to the needs and the intent, will actually help you optimize even further and get to results as quickly as possible.

They could also be different audience segments that are in the same stage of a lifecycle. So being very relevant according to the stage they are in, by using personalization, we’ll also help give them the most relevant experience and optimize their engagement—and ultimately how they convert to the goals that matter most for your organization.

Lead with tests

If you don’t have those specific audiences, then always lead with tests. On top of any personalization you are running, any test you run will help you identify further audiences. Tests could be used with impact, big components on impact pages, and for the site experiences to optimize that. And you can also use machine learning to inform yourself about new audiences that you can turn into personalization.

That’s an example of how machine learning works with optimization in Sitecore to really help you be as efficient as possible to turn visitors into what matters most for your business.

I hope you found this useful and got some inspiration about how you can use testing to inform personalization based on machine learning. If you have any questions or comments, feel free to reach out, or watch any of the other Sitecore sessions. Thank you so much for watching.


The context marketing workbook:
A practical guide for modern marketers

Learn how leading marketers are delivering personal, low-friction experiences—anytime, anyplace.

Learn more