Developing algorithms that let you personalize your website in a granular way using machine learning.
Hi, it's Sander here, CEO at Unless. Today's video is about the difference between rules-based and automated personalization. According to industry research, about 30% of all personalization efforts by marketing teams are highly effective. This sounds pretty good, but it could be improved on. Now, I will tell you how.
So, currently, most marketing teams are doing personalization on their website using rules-based personalization. This means that manually, they create versions of their websites that cater to specific segments of their visitors - which means for example, that if you are a woman, chances are that the website that you see in front of you is specifically tailored to a female audience. This is quite an effective approach. It's very simple to implement, and therefore relatively cheap, especially if you use service providers like Unless. However, the disadvantage is that it is not very granular, because if you are a woman, that is only one side of you, right? So, as an individual, there are way more metrics to take into account. So, if you want to cater to those unique individuals instead - which would be more effective - you should reside to automated personalization.
Automated personalization is a dynamic style of personalization that is dealt with by algorithms, not by people. Those algorithms tend to learn, using machine learning, frombig batches of data and high volumes of traffic. This is immediately the disadvantage, too, because hardly any website owner can actually build on top of such numbers of data. So, while rules-based personalization is relatively simple, automated personalization is more effective. But... it is very hard to implement and therefore more expensive.
There are companies that do this pretty well already. This goes mainly for product suggestions. So, you can think of Netflix, or Spotify, or Amazon who serve content based on who they think their customers are. It would be great if you website would respond the same way, even if you don't have the data to back up an algorithm.
So, at Unless, we started thinking about that and we are currently developing an algorithm that is based on best practices across websites. So, that means that when you use this algorithm, you don't need to have all this data yourself. Secondly, the results of this algorithm will be served to you as suggestions only, which means that you will still be able to keep your rules-based personalization and apply your knowldge of customer segments, but within the scope of that, these automated suggestions will allow you to create a much more granular approach to cater to individuals as well . If you would like to know more about that, feel free to contact us at email@example.com or contact me directly - and I hope to speak to you soon!