Learn About Different Product Recommendation Types
Overview
AfterShip Personalization allows you to choose from 11 product recommendation types to improve your sales figures. Check this guide to learn about all of them.
Algorithm | Description | Store Data Used | Time frame for data accumulation | User Case |
|---|---|---|---|---|
Manual selection | Recommends specific products that merchants want to upsell or cross-sell through manual selection. | None |
|
|
Best sellers | Recommends products that have sold the most based on all historical orders in the store. | Store orders data | 30 days |
|
New arrivals | Recommends products that are new on the shelf. | Products meta information |
|
|
Frequently bought together | Recommends products that are most likely purchased together based on store order history statistics. This works well with bundle upselling. | Store orders data | 180 days | Potato chips and soft drinks |
Complements | Complements offer a subtle way to tie a whole look or function together with the main products, enhancing the customer shopping experience. | Store user orders, products meta information | 180 days | Phone case for a phone; replacement bags for a vacuum |
Similar products | Recommends products that are semantically or visually similar, without using any customer data. | Products meta information |
| Pink t-shirt and pink long sleeve t-shirt; sandals and slippers |
Substitutes | Substitute products are alternate items that can be used in place of a desired item. They have the same features as the main product but may differ in price, availability, or quality. | Products meta information |
| Substitute wine with grape juice; substitute a low-priced handbag with a more expensive one |
Same product upsell | Recommends the most expensive or cheapest products in the current cart or order, usually combined with a discount. | User cart and orders data | 1 day |
|
Free shipping upsell | Recommends products that could help reach the free shipping threshold when purchased. | User order and product price data | 30 days |
|
Trending | Recommends products that are currently trending in the store based on user views or orders within a few days. | Store clicked data | 30 days |
|
NextLLM | Real-time recommendations for the next best product series using a self-training multifaceted predictive model based on product tagging data and shopping behaviors from all AfterShip’s stores. | Products meta information, store orders data, and user clicked data | 30 days |
|
Complete the look | Recommends different types of products that visually pair with a target product. | Products meta information |
| Shoes for a dress; rings for a bracelet |
Recently viewed | Recommend products that the customer has viewed but not purchased. If applied on the product detail page, the currently viewed product will be excluded from the recommendations. Similarly, if used on the cart page, cart drawer, or smart cart, any products already in the cart will be excluded. | User clicked data | 30 Days |
|
Updated on: 07/03/2025