Image from: Sweet Savings
Based on the e-mail recommendations I receive, I’m pretty sure Amazon thinks I’m juggling school with being a new father, and that I enjoy camping, makeup, and video games (only some of which are true). But I’m willing to bet a lot of this is skewed based on the fact that I work on the retail team at a web analytics company and that sometimes has me browsing prestige beauty products for hours on end (this is for work, I swear). So while I generally ignore the e-mail recommendations I get from Amazon every few days, knowing that it’s probably for something I have little interest in. I was wondering whether people with normal browsing behavior actually take a look at these recommendations and how good Amazon is at guessing what they want.
If you do click the link to see your recommendations in that e-mail, you’ll be taken to “Your” Amazon.com (you can also navigate there from the homepage without the email). On that page, you’ll find all types of recommendations, all based on your browsing, purchasing, and rating history (Note: this is separate from the “Customers Who Bought This Item Also Bought” section on a product page based on the what you’re currently viewing). The first thing I wanted to know is how many people actually go to “their” Amazon.com. The answer is that in the past year, an average of just of 5% of monthly visits included a stop at this page. In fact, visitation was lowest this past fall, but has been making a slow comeback since, although it is still below visitation during this time last year.
So we know that not too many people actually visit “their” Amazon.com, but how successful are Amazon’s recommendations for those shoppers that do visit? Looking at how many people add a product Amazon has recommended to them to their cart, we see an average of 5.31% recommended product viewing visits have that product added to the cart. When we compare this to the general Amazon browsing population, it is just over half of the average 9.13% of product viewing visits that include an add to cart. It seems that Amazon recommendations are just not making their way into the cart as often as other products.
Well, of the people that are drawn into Amazon’s recommendation page, they’re not adding too much to their carts from there. But could it be possible that while not much is being put in shopping carts, they could still be less likely to abandon their purchase in the cart because Amazon’s recommendations require no second guessing? On average, fewer than 2% of recommended product views end up checking out. Looking at the general Amazon browsing population, more than 9% of product views result in a competed checkout. So checkouts are much more likely to be driven from ordinary product views than from recommended products.
While one possible conclusion to draw from this data is that Amazon just isn’t very good at predicting the products shoppers are looking to buy, it may actually say more on how people shop on Amazon. People may be going to Amazon with a specific need in mind, rather than to browse what the site has to offer. Whenever I shop on Amazon myself, it’s nearly exclusively with a product already in mind, rather than just “browsing the aisles” (which in all seriousness could take you days on Amazon). So if you are like me and don’t generally look at what Amazon’s been recommending to you, maybe you should check out “your” Amazon.com next time you end up there and see for yourself whether Amazon knows you a little better than you thought.