Area: Optimizely Product Recommendations, Optimizely Personalized Search & Navigation, Optimizely Content Recommendations
Applies to versions: Not applicable

Install and configure personalization

Recommended reading 

This topic describes the NuGet packages and initial configuration for the setup of product recommendations and personalized search results with the Optimizely Personalization suite. To get personalization running for a site, configure access to the tracking API to post usage data and get recommendations back.

In this topic


To install and configure, you need the configuration keys and credentials provided by Optimizely during onboarding of the service. Contact Optimizely for questions regarding the onboarding process.

  • See Installing Optimizely to set up Visual Studio with Optimizely extensions and the NuGet feed source.
  • See Content recommendations to implement Episerver Content Recommendations.
  • See below for NuGet packages needed for implementing product and email recommendations, and personalized search results.


In addition to onboarding the service, install the following NuGet packages for the tracking and personalization features:  

  • EPiServer.Find.Personalization. Native integration for search recommendations (Personalized Search & Navigation).
  • EPiServer.Personalization. Integrates Personalization Portal UI (Product Recommendations and Email Product Recommendations) with the Optimizely top menu.
  • EPiServer.Personalization.Commerce. Native integration for product recommendations (formerly Perform and Reach).
  • EPiServer.Personalization.Common. Common base API classes.
  • EPiServer.Profiles.Client. Main package for tracking of an Optimizely site to Profile Store.
  • EPiServer.Profiles.Client.Common. Contains common classes used by other packages.
  • EPiServer.Tracking.Cms. CMS integration for tracking of data.
  • EPiServer.Tracking.Commerce. Commerce integration for tracking (Product Recommendations and Email Product Recommendations, Profile Store, Personalized Search & Navigation).
  • EPiServer.Tracking.Core. Generic package for tracking.
  • EPiServer.Tracking.PageView. Native integration for tracking of CMS/content data to Profile Store.


Native integrations

Installing a main native integration package also installs other packages based on dependencies. The package sets provide product-specific access to the tracking APIs:

  • Product recommendations for Optimizely Commerce
    • Optimizely Product Recommendations and Email Product Recommendations tracking and recommendations: EPiServer.Personalization.Commerce.
    • Optimizely Personalization Portal integration: EPiServer.Personalization.
    • See Installing and configuring the native integration package for Commerce.
  • Personalized search results for Optimizely Search & Navigation (formerly Find)
    • Optimizely Personalized Search & Navigation: EPiServer.Find.Personalization.
    • See Personalized Search & Navigation to learn how to install and configure this native integration.

General integration APIs

Instead of using the native integrations, you can use the Server-to-Server API and JavaScript API to call the HTTP endpoints directly to integrate with the core personalization tracking API, and send and receive tracking data and recommendations in JSON format

Importing data

For Optimizely Commerce sites, the core part of the personalization functionality consists of a catalog product feed that provides the input data, combined with a tracking script that enables Optimizely to track behavior and serve suitable recommendations. When you set up the personalization, you configure the feed and import the source data. See Data import guides.

Recommendation algorithms learn which products to recommend. 

  • If you see large volumes of orders per week, the system should learn quickly, and start recommending highly relevant products.
  • If you take limited orders, or you have an incredibly expansive product catalogue, it will take longer for the system to learn as there is less input relative to an extensive inventory.

 You should import historic orders prior to launch; see Offline and historical orders

Additionally after launch, you can continue to send offline orders through as an XML file or as a real-time API reporting (if you can hook that up to your offline order processing system), if their volume is pertinent for your business, and the offline orders would help the algorithms to produce optimized results. See Offline order tracking

Personalization API

Customize personalization with the Personalization API. This API contains header information, endpoints, and response messages for tracking and recommendations. The client calls the Personalization API that (primarily) returns product or content recommendations, to be used in various output channels.  See Personalization API.

Profile Store and Visitor Intelligence

Using Optimizely Profile Store and Visitor Intelligence depends on how personalization is set up for your solution.

  • Optimizely Profile Store, also known as the Optimizely Customer Data Management Platform, stores tracking data, which then can be processed and used for visualizing data, and segmenting visitor profiles.
  • Optimimizely Visitor Intelligence is the user interface for vizualizing data and segmenting profiles.

See Visitor Intelligence and Personalization to learn how to set up tracking clients and the profile management user interface.

Feature-specific configurations

Below are some product/feature-specific configurations. 

Email Product Recommendations

Contact Optimizely to provide the mapping of email addresses to pseudonymized IDs (enabling use of hashed IDs when using Optimizely Campaign), and Optimizely will configure this in the Personalization Portal.

Product Recommendations

Contact Optimizely to configure these features:

  • Allow updated orders to affect personalization.
  • Multi-site support for recommendations.

Content Recommendations

Contact Optimizely to configure availability in Optimizely DXP regions - East/West US, EMEA, APAC.

Related topics

Do you find this information helpful? Please log in to provide feedback.

Last updated: Jan 31, 2019

Recommended reading