I don’t think of myself as an acquisitive person, but recently, I’ve been shopping a lot. A lot, a lot. Birthday presents, baby presents, graduation gifts – for college and high school grads – and house renovation necessities have turned me into a full-time shopper, it seems.
I’ve written about those experiences and how retailers could have helped me shop by using context-based recommendations, which rely on social proofs, and business-based recommendations, which focus on the retailer’s goals.
These are part of our ebook, The Periodic Table of Recommendations, in which we explore how retailers group recommendation types together to help shoppers find what they need – sometimes even before they realize they need it.
What Are Profile-based Recommendations?
Profile-based recommendations suggest products based on a shopper’s profile and preferences – and the more you know about a shopper, the better the recommendations can be. More importantly to you, these recommendations can help you increase conversions.
I’ve been renovating an American Foursquare built at the very beginning of the Arts and Crafts era in 1902. It’s a house with good bones and in need of a lot of care. I’ve worked with a large American retailer to equip it and am now working on furnishing it. My central question is: Can I make it comfortable while honoring its past?
My retailer knows this, has seen pictures of the house, understands my goals, and will probably send me a birthday card. So I think we can use this as an illustrative example of how profile-based recommendations can work. If you read my last post, you’ll know that this retailer doesn’t have a recommendation engine, and that’s all the more reason to talk about it here.
How Do Profile-based Recommendations Work?
Profile-based recommendations are going to sound familiar because we see them all over – they are buy-it-again, recently viewed, and top picks (sometimes called affinity-based recommendations). What’s new is how to use these recommendations to increase average order value and conversions – and ultimately, the bottom line.
Profile-based recommendations use a plethora of data that customers voluntarily give you and combine it with data you gather while they’re on your site. These include:
- Explicit data (provided through surveys and registration forms);
- Intent data (e.g. need, want)
- Offsite behavioral data (e.g. search, paid media, publishers)
- Onsite behavioral data (e.g. site visits, search queries product page views)
- Demographics and personal data (e.g. interests)
- Implicit preferences (e.g. brands inferred from clickstream)
These combine to create highly personalized recommendations that save shoppers valuable time. For my house project, they would include suggestions for period-reproduction cabinet pulls, genuine antique lights or reproductions, and custom-sized doors, for example.
This kind of hyper-personalization differs from personalization, which is knowing who someone is when they get to your site. Hyper-personalization, or 1:1 personalization, is knowing what they want when they get there.
This is particularly tricky since personalization is in the eye of the beholder, which means that ecommerce sites need to be flexible enough to provide a personalized site for every shopper who visits.
Profile-Based Recommendations Speak Directly to the Consumer
Buy-it-again recommendations are reminders to re-purchase frequently-used consumables. Base these on analysis of prior purchases to create repeat shoppers. Since I’m going to be renting my property part-time, I would like buy-it-again notifications for laundry soap; paper towels; sheets; and items I’ll use as welcome gifts, such as peach salsa.
Recently viewed or view-it-again recommendations are a light form of personalization that suits both new and returning customers. They can help customers avoid frustration by helping shoppers easily find an item they recently looked at, offering reminders that make it easy to compare products, or going back to an earlier choice to complete the sale.
During my renovation, I used these at least 50 times and wished for them probably 100 more. I bookmarked pages on sites that didn’t have this functionality. I also combed through my viewing history. Probably half of the time, I couldn’t find the product again when I was ready to buy because I misremembered the name or for a reason I never figured out. Those retailers left my money on the table.
You can put these recommendations pretty much anywhere on your website, but you want to do it before shoppers check out, ideally on a product detail page.
I often have only an idea of what I want to buy – and I’m surely not the only one. Personalized recommendations based on preferences, interests, and needs can be an incredibly valuable tool for inspiration and product discovery.
We call “top picks for you” and “recommended for you” affinity-based recommendations, and they draw on shoppers’ profiles, interests, preferences, what they’ve looked at on your site, and previous purchases to find products that are of interest to individual shoppers. Think of these as offering your customers a personal shopper. In my case, I would have loved to see products that matched other things I was buying, such as towel bars in the same finish and style as the shower trim.
In addition to impacting your conversion rate, affinity-type recommendations can impact your clickthru rate as well.
Machine Learning Helps Retailers Nudge Customers
Sometimes, a shopper just needs a little nudge. Using profile-based recommendations can provide that nudge by putting items they want – but might not know to search for – in front of them. These recommendations make it easy for them to buy.
Profile-based recommendations can be implemented and executed in quite different ways — leveraging different sorts of data — but fundamentally they use customer data to cut through the noise, increasing relevance and speeding customer access to desired functionality and content.
They can be used across different stages and touchpoints through the customer journey, based on shoppers’ profiles, interests most notably in Homepages and No Results Pages.
Machine learning and other artificial intelligence techniques can help corral the huge amounts of data shoppers create as they shop – allowing retailers to put hyper-personalized pages in front of every shopper so that they see what they want with as little effort as possible.
This personalization also benefits ecommerce companies because they can promote their most profitable products without having to spend endless hours tweaking rules that help bring those products to the forefront of their sites.
Dig Deeper
Investigate all of the recommendation types, and what formula works best for your ecommerce site.