A one-size-fits-all digital marketing strategy doesn’t work. That’s why companies are employing personalization on their websites and in their apps. But personalization isn’t one-size-fits-all either! It comes in many forms.
In this second part of a two-part blog series, I take a look at algorithmic personalization, and I include B2C Retail and B2B Technology company examples to help explain. In the first article of the series, I examined rule-based personalization.
The Personalization Spectrum - Algorithmic Personalization
Algorithmic personalization is an automated form of personalization that utilizes machine learning and predictive analytics to present the most relevant content or experience for each and every visitor. This is particularly effective for product and/or article recommendations, and is ideal for organizations that are under-engaging their users or looking to optimize conversion rates, time on site and revenue per customer.
General recommendations are made based on what others like the current visitor have browsed, read or bought previously. Continuing with the examples from Part 1, the retailer using general recommendations might recognize the regular visitor from Florida and the fact that she’s spent a lot of time looking at swimwear, and recommend the swimwear items on sale. The B2B tech vendor should recognize a repeat visitor and the fact that he’s shown a lot of interest in network servers (and perhaps has downloaded the promoted white paper), so now it would be appropriate to recommend customer testimonials or blog posts based on his interests.
For truly individualized recommendations, marketers can take everything they’ve learned about each visitor and even factor in their intent in real time. For our online retailer, the shopper has shown interest in specific styles of swimwear and perhaps even particular brands. Not only would the retailer want to highlight the specific products she’s spent the most time on, but it should also recommend the best-selling swimwear items in the styles and brands she likes most. For the B2B site, if the customer has checked out various content and blog articles about network servers, the company could make individualized recommendations based on the most popular posts on that topic – that the person hasn’t read yet – and insert dynamic calls-to-action (e.g. “speak to a rep about our network servers”) into the articles he’s reading.
As you can see from the examples, there’s a time and a place for different types of personalization, whether they’re rule-based or algorithmic, broad-based or individualized. By running iterative tests and having an environment where you can implement a variety of techniques, you can discover the types of personalization that work best for your visitors and align to your website goals.