You navigate to an e-commerce site looking to buy a t-shirt. After adding the t-shirt to your shopping cart, the e-commerce stores "personalization" algorithm decides that many people who bought that type of a t-shirt also bought a pair of jeans. So you get recommended multiple different pairs of jeans. Great experience, right? Wrong.
The reason you were buying that t-shirt in the first place was that you were going to spend the next three weeks in Hawaii with your two children, tanning in the sun and enjoying the Aloha-lifestyle.
As a result, there is a 0% chance you're interested in the pair of jeans the retailers "personalization" algorithm decided you wanted. You were just placed into a bucket of people "like you" in "behaviour group A". Starts to feel like an awfully impersonal experience, doesn't it?
You decide to abandon your shopping cart and move to the next e-commerce store. This e-commerce store uses real 1:1 personalization. They know based on external information about your plans, lifestyle, social media posts and other information that you're planning that trip to Hawaii before you even place that t-shirt into your shopping cart.
When you do, in stead of showing you a pair of jeans that "shoppers in behaviour group A are likely to buy", this retailer recommends you would consider some sunscreen lotion and a fresh pair of shorts. Based on external information this retailer also knows you have two children and recommends you buy some beach toys for them as well as some UV-protected swimming wear. You are impressed that the retailer understands not just what you are looking for but WHY you're looking for it and decide to buy all of the recommended items.
The resulting increase in sales versus abandoned shopping cart is the difference between treating your shoppers as individuals as opposed to bucketing them into "behaviour groups" in your e-commerce stores.