LS Recommend is a module in LS Central developed by LS Retail to help clients predict what their customers want and to enhance product discovery.
It is a cloud-based Software as a Service (SaaS) solution based on Microsoft Azure Cortana Intelligence.
LS Recommend enables clients to create a recommendation model based on past customer activity, which can in return provide customized product recommendations. The product recommendations are consumed by LS Central and LS Commerce and can be displayed on mobile, stationary and web POS, e-commerce site, and loyalty app.
The models created by LS Recommend learn from usage data that correspond to a catalog of items to predict which items are more likely to be of interest to the user. Recommendations are based on an item or a list of items, and optionally on a user ID in order to get personalized recommendations for that user. The following scenarios are supported by LS Recommend:
- Item to item: Recommend items to a single selected item from the basket.
- Item to basket: Recommend items that are likely to be consumed together, given an item or a collection of items in the shopping basket.
- Item to member or customer: Return personalized recommendations based on the member's or customer's purchase history.
- Item to basket with member or customer: Combine options 2 and 3 by recommending items based on items in the shopping basket, and on a member's or customer's purchase history.
If no purchase history exists for a member or a customer, then a recommendation is automatically based on items in the shopping basket. Empty recommendations may be returned if none of the items are in the catalog or if the supplied data was not sufficient to give a meaningful recommendation.
Key concepts
|
ModelA model is the core component of LS Recommend. A model can be viewed as a container for the input data, parameter settings, and the underlying mathematical model which makes the product recommendations. Only one model can be in use to give recommendations at each time in LS Recommend. Each model can have multiple versions which are referred to as builds.DataAll models provided by LS Recommend learn from a given data set in order to provide product recommendations. The data supplied to a model consists of catalog information and usage history.
BuildA build of a model is created when data files have been successfully loaded to the model container. This procedure is called “creating a model build” or just “creating a build” which starts the training process of the machine learning algorithm that generates a build. After a build is created the model is ready to give recommendations. |
Using the LS Recommend help
If you want to get started immediately with LS Recommend default settings, follow the steps in Quick Start.
Detailed information on relevant procedures and model components is provided in the help topics. To make the navigation easier, links to specific sections that describe useful tasks are listed in the following table.
| To | See |
|---|---|
| Get started with default settings. | Quick Start |
| Set up LS Recommend. | Setup |
| Create a recommendation model. | Models |
| Trigger a new build for an existing model. | Builds |
| Include features in a recommendation model. | Features |
| Block or promote items that appear in the recommendation result. | Result Filters |
| Apply display rules to control which products are displayed for recommendations in POS. | Display Rules for POS |
| Automate model processes. | Scheduler Jobs |
| Display product recommendations in LS Commerce. | LS Recommend in Commerce |