Personalization is everything right now for ecommerce shops. When your customer feels like you know them and understand what they want, they are more likely to buy from you again (and more often). In fact, an Accenture report says personalization increases the likelihood of a prospect purchasing from you by 75 percent, and McKinsey found that 35 percent of sales on Amazon and 75 percent of Netflix sales were directly related to product recommendations.
There are a ton of ways to personalize your Shopify store, but a great place to start is with recommending products to your customers.
Once someone is on your site and actively looking at a product, you already know something about them. They’re interested in that product. Now, maybe it doesn’t turn out to be what they want, or they change their mind about what they want, or the price is to high, or whatever. But you still know that item peaked their interest. Using that one simple piece of data, there are tools that can help you suggest other similar or complementary products that they may also like.
Product recommendations are a powerful tool. They have been shown to:
Lift conversion rates
Deliver an improved user experience by acting as a filter to help shoppers quickly get to what they want to see
Bring in more return shoppers
Boost the average order value
Increase conversion rates
BOOM. Your shoppers feel like you’re paying attention to what they want, and you might just suggest a product that they love and buy.
Creating On-Target Product Recommendations [The Algorithm]
The technology that powers product recommendations is called an “e-commerce recommendation engine,” and it uses machine-learning algorithms and complex mathematical principles. Huh? This is the behind-the-scenes stuff. You don’t need to know how to build it, you just need to know how to make it work for your site.
At its most basic, an e-commerce recommendation engine targets the right products to customers by learning about a customer’s preferences (including their level of buyer intent and shopping habits) and combining that with data collected from other buyers who display similar traits. That data informs personalized recommendations that improve with every interaction.
You have the ability to configure the recommendations in several ways:
Behavior-Based Filtering: Collects data about every individual visitor over multiple visits (including viewing history and buying behavior), and makes recommendations based on that person’s past choices.
Collaborative Filtering: Collects data from shoppers who have made similar choices to make recommendations. This approach also incorporates data from best sellers.
Hybrid: Combines the behavior-based and collaborative-based methods, but focuses the recommendations on attributes of the specific, individual shopper.
Manual Curation: No algorithm needed - you can manually select products based on internal preference, such as items for liquidation.
Recommended Pairings: Define sets, bundles, etc. to be promoted together.
Where to Promote Recommended Products
Should you have product recommendations in the cart and on the product page? What about on the homepage? Anywhere else? Yes and yes. Product recommendations work best when they’re promoted at multiple touch-points, through different pages on your site and across a variety of pre- and post-sale marketing channels. Here are some locations to consider:
Recommendations on the product pages offer a “next step” in the shopper’s search to keep them browsing your site. The more time they spend in your shop, the higher the chance they’ll make a purchase. The strategy here is to display the most relevant items; either alternatives or complimentary products to the one being viewed.
The customer has already decided to make a purchase, so they’re probably more willing to say yes to other offers. Recommending products that are similar of related to those in the shopper’s cart can boost average order quantities and values. According to a study of 300 ecommerce stores by Braillance and Marketingsherpa, product recommendations in the cart were among the 10 best performing recommendation types.
Visitors who start at the homepage don’t necessarily come looking for anything specific, so recommendations here should inform visitors about current deals and discounts, and showcase your product portfolio with a focus on popular products and curated collections.
Email marketing can reinforce product recommendations and encourage shoppers to return to purchase a recommendation or can deliver additional suggestions. Standard transactional emails, like shipping notifications, can also include personalized product recommendations.
When a customer is looking at a collection page, you already have valuable information about what they are looking for. The goal here is to help your customer find their desired product quickly and easily. While the best tool here is often advanced filtering and search, product recommendations can also provide value.
8 Ways to Promote Product Recommendations (On a Scale from “Gentle Suggestion” to “In Your Face”)
Sometimes, you want to whisper gently in your shopper’s ear, “It’s possible you may be interested in this as well.” Other times, you’re in a position to say, “I know what you’re looking for, and you’re going to love this. Here, try it.”
Here are a few different ways to consider using product recommendations in your Shopify store - starting with more gentle suggestions and working our way up to more direct, “in your face” recommendations:
1. Popular Products: Based on the number of times an item has been purchased, this approach helps shoppers find what they’re looking for by showing them what most people are looking for. More sophisticated systems may incorporate clicks, add-to-carts, and other data.
2. Rating-Based Recommendations: It’s usually too difficult to translate comments or ratings into data that can be used in a recommendation engine, so ratings-based recommendations are instead based on implicit feedback like clicks, purchases, and browsing habits to suggest “Best Sellers”.
3. Related Products: Simple category-based filtering can be done even without a recommendation engine, although the algorithm can be useful to include metadata like product descriptions, product titles, tags, and prices for more granular filtering. For example, this approach could recommend other items of the same brand or color.
4. Personalized Recommendations: Display different product recommendations to each visitor based on their past purchase and browsing history. This may look like, “Since you already own this, you may also want this…” These types of recommendations require considerable amounts of behavioural data on users.
5. “Customer Who Bought/Viewed This Also Bought/Viewed…” Collaborative Filtering: This type of recommendation combines preference data from many users with analysis of the similarity of two products (by look at how often they/re present together in browsing or purchase histories). Together, this data can fuel both item-to-item (similarity-based) recommendations as well as personalized recommendations.
6. Cross-Selling: When items are added to the cart, a recommendation engine can suggest complementary accessories (ex. Socks for a shoe purchase, batteries for an electronic toy, or a belt to go with a dress). It’s very hard to automate the process of recommending compatible accessories. Manually assigning complementary accessories can be an administration-heavy approach, however the payoff can significantly increase average order size and value on your site.
7. Frequently Bought Together: Another data-heavy technique, but one that can be worth the time to increase the average order value. Items that are frequently purchased together can be promoted as a bundle on the cart page (possibly with a discount for buying the bundle).
8. Upsell: Product recommendations are used to encourage the buyer to move up to a more fully-featured version of the product they’re currently considering.
Granted there is a deluge of applications and widgets out there that can help you create the kind of recommendation engine that’s right for your store, but the key to optimizing your recommendations engine is that same as with most e-commerce configurations: test, test, test.
Need help getting started with a product recommendation engine, or optimizing your results? Growth Spark is a Shopify Plus expert in design and development, and they would be thrilled to help you grow your store with a product recommendation engine and configuration that will get you results.