Predictive Product Bundling

Who can use this feature?

This feature is available to any Athos Commerce customer as an add-on to your existing plan. If you're interested in taking advantage of it, please reach out to your customer success manager.

Note that product bundling does not apply any additional discounts to the featured products. The purpose of bundling is to help shoppers discover more products and quickly add them to their carts. If the products already have a discount applied, that discount will display in the bundle; the bundle itself is not discounted further.

You can see detailed analysis of bundling use in your store via the Bundle Recommendations report.

Introduction

Merchandisers now have the ability to automatically present bundling options for items that users often purchase together. When a shopper is looking at the product detail page for an item, an optional bundle can be displayed near the "Add to Cart" button, and it will show the current item (the "seed" product) with additional items that can all be added to the cart with a single click. In the example shown below, the orange sweatpants are the seed product, shown with two complementary products to create the bundle.

Below is an example of how this might be configured in the Athos console, with the bundle profile filtered by ss_product_type of "apparel":

The Recommendations API is aware of the exclusion status for each product, so if a bundle is requested and the seed product is excluded, then no results will be returned. This prevents the store from showing a bundle that cannot be added to the cart.

The recommendation service already prevents excluded items from being recommended, but it doesn't block them from being a seed item. That is by design, because it is valuable to show recommendations (especially similar items) when the shopper is looking at the product detail page (PDP) for an item that is not available.

Note: This feature only affects bundles, since bundles are cross-sell recommendations with the seed product included as the first item.

Setup

Setup is simple. In the Athos console, navigate to "Personalized Recommendations" and click "Add New Profile".

Select "Product Detail Page" and click the "Next" button in the bottom right corner.

Select the bundle type by clicking on the preferred type. Options include the following:

• Session Bundle = Recommend products based on the individual shopper's recent activity
• Cross-sell Bundle = Recommend products that go well with the product being viewed
• Cart Cross-sell Bundle = Recommend products that complement items that are already in the shopper's cart
• Trending Bundle = Recommend popular product bundles based on site, category, or brand trends
• Custom Bundle = Curate a filtered list of products based on the criteria of your choice

Click the "Next" button in the bottom right corner. In the Bundled recommendations page, the Profile Editor will appear.

Click the "Template" drop-down menu to select the bundle template.

Maximum Bundle Size

Set the maximum bundle size (number of products) in the field. This value represents the maximum bundle size — including the seed product. The default setting for the maximum bundle size is three (3) = the seed product + 2, but users can set the bundle size between two and 10. The minimum size is two (2) = the seed product + 1. The recommendation service uses the precision setting to restrict results; it will return only results that meet that setting, so the actual bundle size will vary depending on the seed product used in the request.

Title

Now you can configure the name of your bundle profile in the Title field; you may select from a default template or from custom templates (if these have been configured). 

Filters tab

Use the Filters tab to limit which products are allowed to appear in the bundle. The fields used here must first be enabled in the Recs Filter column on the Field Settings page.

Note: Adding filters to a profile may result in fewer results. This may cause the profile not to appear on all pages.

Click the "Add Filter + " button to create a new filter.

Click the "Choose" drop-down menu to select a filter. Then click the selector button to determine the formula: "is", "is not", etc.

In the example shown above, the filter is for "product_type", the selector is set to "is", and the product type is "BEDDING".

When you have finished configuring the filter, click the green checkmark to confirm it.

Configure any other filters you would like to apply, or click the "x" at the right of the filter definition to remove the filter. When you are done, click the "Save Profile" button at the bottom right of the window to save your changes and close the Profile editor.

Precision tab

Click the Precision tab to configure how specific recommendations are. This value dictates how many conditions the system uses to find recommended products. Broad precision will expand the recommendations algorithm to ensure that products can be recommended even when shopper data is sparse. Narrow precision will restrict recommendations to prevent the algorithm from defaulting to related products or top sellers. This could result in bundles not being displayed for certain seed products.

Precision selector

Drag the slider to select the precision. As the precision gets broader, it includes all recommendations from narrower precision settings. Note: Depending on the recommendations type you have chosen, some of the options below may not be available. In the example above for cross-sell bundle recommendations, the "Trending" option is not available.

• Curated = Curated recommendations will be limited to the products set in Product Rules. If the seed product does not have one or more products associated with it in a rule, no recommendations will be returned. For example, if merchandisers want to create a "Complete the Look" experience, they would set the precision filter to "Curated" so the recommendations carousel would only show the pinned products.
• Cross-sell = Cross-sell recommendations are used on a page to recommend products that go well with the product that the shopper is viewing. 
• Similar = Similar recommendations are products that a shopper should consider buying as an alternative to the product they’re currently viewing.
• Trending = Trending recommendations are popular product bundles across a rolling window of time. This recommendation type supports either site-wide trending or trending within a category or brand.

Integration instructions

At the bottom left of the Profile editor, no matter which tab you have open, the system will display the "Show Integration Instructions" button. Click this button to see integration instructions for your particular installation.

A new window will pop up and will display a link to open instructions specific to your installation: Snap, API, etc.

Click the "x" in the top right corner to close this pop-up window.