15 Product Recommendation Models to Enhance Your eCommerce Homepage
We often receive questions about the best product recommendation model for website visitors. Our answer? It depends on the context...
About the Author:
Co-founder of Datacop, agency that fulfils marketing operation roles in large eCommerce companies such as OluKai, Melin, Roark, Visual Comfort and Company, Dedoles and others.
We often receive questions about the best product recommendation model for website visitors. Our answer? It depends on the context.
Consider a first-time visitor versus a returning customer who has already made two purchases. Should they receive the same recommendations? We believe the answer to these questions is no.
Different scenarios demand different strategies. For some, showcasing top-selling products may suffice. For others, more sophisticated models like collaborative filtering could prove more beneficial.
In this article, we'll explore a variety of product recommendation models and discuss the specific contexts where each can be most effectively applied.
First-Time Visitors
First-time visitors present a unique challenge due to the limited information available about them. They have no previous purchasing history or demonstrated preferences, making it difficult to predict what products might interest them.
For these visitors, we recommend a broader approach to product recommendations on the homepage. Instead of showcasing specific products, we believe it’s more effective to highlight product categories. This gives newcomers a comprehensive view of what your eCommerce site offers.
If your store features a wide range of product categories, focusing on the most popular ones can be particularly effective.
For instance, a furniture store could inspire first-time visitors by emphasizing its broad assortment rather than specific items, providing a welcoming and informative introduction to what’s available.
If you're considering displaying products on the homepage for first-time visitors, it's best to feature items that are most likely to capture the attention of a broad audience. Consider the following recommendation models:
1. Trending Products
Trending products are generally the bestsellers, though they come with certain limitations. Alternatively, you might feature products with the highest conversion rates or the best profit-per-visit metric. For further insights on this topic, refer to our detailed article.
2. Best Deals
If discounts are part of your eCommerce strategy, showcasing the most discounted products on the homepage can effectively attract first-time visitors.
3. New Arrivals
Simply feature the latest products added to your inventory. This approach keeps your offerings fresh and engaging.
4. Manual Selection
Manual selection is a straightforward yet potentially effective product recommendation strategy. This approach is ideal for promoting products during special events, such as Mother's Day, or for moving excess inventory. By incorporating these items into your manual recommendation model, you can increase their exposure and, consequently, their sales.
Returning Visitors
Returning visitors provide more personalization opportunities than first-time visitors because we have access to data about their past interactions. This data includes viewed products, items added to the cart, their last visit, and categories they've shown interest in.
5. Recently Viewed Products
Display products that customers have previously interacted with on the homepage to allow them to easily resume their browsing.
6. Left in Your Cart
Highlight products that were added to the cart but not purchased. These items likely hold a stronger interest for the customer since they were closer to making a purchase.
7. New Arrivals Since Your Last Visit
This model tailors the "New Arrivals" strategy with a small twist by leveraging data on when returning visitors last accessed the site. It displays products that have been added to the store since their last visit, providing a personalized touch that enhances the shopping experience.
8. Trending Products in [Category]
"Trending Products in [Category]" is a personalized twist on the general "Trending Products" model discussed earlier. For instance, if a customer focused on the "Living Room" category during their previous visit, we can highlight trending products from this category upon their return.
This strategy is likely to be more effective, as it aligns with the customer's demonstrated interest, enhancing the potential for conversion.
9. Discover
If a customer interacts exclusively with specific products or categories, continually recommending similar items might limit their exposure to your broader catalog. This is where the "Discover" recommendation model becomes valuable. It introduces products from categories the customer has not yet explored, encouraging them to discover new selections. This approach not only diversifies their shopping experience but also broadens their engagement with your store's full range of offerings.
10. Similar Products
Consider a customer who primarily interacted with black shoes during their visits to your store. You could recommend similar products to enhance their shopping experience. There are two approaches you might consider: first, you could suggest products that share the same or similar attributes, such as other black shoes. Alternatively, you could employ a collaborative filtering technique to recommend products favored by other customers with similar browsing habits.
Returning Customers
Returning customers, who have previously made purchases with us, represent a group for which we have a significant amount of information. While all the recommendation models discussed in the "Returning Visitor" section apply to them, there are two additional models particularly relevant for this group:
11. Buy Again
I regularly purchase protein powder from a major local retailer, consistently choosing the same product. Ideally, this item should be prominently displayed on the homepage for easy reordering upon my return to the site. Unfortunately, this retailer does not currently utilize this approach, which suggests a potential area for improvement that could enhance customer convenience.
12. You Might Also Like
This recommendation model leverages a customer's past purchase history to suggest additional products they might find appealing. We can employ several strategies to ensure these recommendations are relevant and enticing:
Similar Products: For customers dedicated to a particular style or type, such as those only purchasing curved-bill premium hats, we recommend similar items within the same category to maintain consistency with their known preferences.
Complementary Products: Here we suggest items that pair well together. For example, products featuring matching ornaments can be grouped to appeal to customers looking for aesthetically cohesive options.
Collaborative filtering: Let’s say a specific customer bought Product A. Collaborative filtering then studies the purchase behavior of customers who also bought Product A. It then analyzes which other products were commonly purchased by this cohort of customers. Based on this analysis, the recommendation model suggests these related products to the original customer who bought Product A, aiming to reflect similar preferences.
Based on Customer Preference (Zero-Party Data)
In this section, we explore several product recommendation strategies tailored to individuals who have explicitly shared their preferences with us. These models can be effectively applied at any point in the customer's journey. For instance, if a customer indicates their preferences during their initial visit, we can immediately personalize their experience. This approach is equally beneficial for engaging our most valuable customers.
One method to gather customer preferences is through microsurveys on our website. Feel free to download our free templates included in this article to implement this strategy effectively.
13. In Your Size
For example, if we know a customer's shoe size is 44, we can specifically recommend products available in that size, ensuring a more relevant and streamlined shopping experience.
14. In Your Favourite Style/Color
When we know a customer prefers specific styles or colors, such as black shoes, we can tailor our recommendations accordingly. This ensures that the products we suggest align closely with their established preferences.
15. Back-in-Stock
When customers request notifications for products that are out of stock, consider creating a specialized recommendation model that is shown to them once these items are available again. Additionally, it may be beneficial to expand these recommendations to include other products they've viewed that have also returned to stock, even if they haven't explicitly requested notifications for these items. This approach has significantly increased revenue in our back-in-stock email campaigns. For a more detailed exploration of this strategy, you can read about the use-case here.
If you found this post valuable…
We hope you found this article valuable. If so, please consider subscribing (for free!) to receive updates on our latest publications.
Additionally, if you think someone else might benefit from this information, we would greatly appreciate it if you could share this article using the button below.