Search the Omeda Knowledge Base

< All Topics

Content Recommendations

Omeda’s Content Recommendations feature includes powerful data science algorithms that enable users to recommend links to their page visitors . The Content Recommendations algorithm will run nightly across the audience members who have visited a page in the last 15 days and uses collaborative filtering to place those audience members into different cohorts. Page recommendations for each audience member will then be created with page recommendations by the strength of the recommendation. These recommendations can then be displayed to users on your site within a personalization so they can easily navigate to pages they are likely to be interested in.


Enabling Content Recommendations

In order to configure Content Recommendations, a user must first have access to Content Recommendations and Olytics. To discuss adding this tool to your Omeda portal, reach out to your Client Success team member.

Once access is given, users can configure Content Recommendations to run on individual Olytics behaviors via the Olytics Integration screen (Menu > Activate > Integrations > Olytics).

In the Manage Domains and Behaviors tab, select Edit beside an Olytics Behavior. In the Edit Behavior modal, check the box to Enable Content Recommendations.

After Content Recommendations are enabled, the behavior will run in the nightly Content Recommendations job later that night.

Excluding Pages from Content Recommendations

By default, all pages visited by your audience in the last 15 days will be pulled into the Content Recommendations nightly job. However, you may want the recommender to exclude certain pages, like your homepage or author descriptions, from being recommended. You can do this via three different methods.

Excluding by Page Type

In the Manage Content Recommendations tab on the Olytics screen, a drop down will appear containing all of the Olytics Behaviors that have Content Recommendations enabled. To add an Exclusion by Page Type, select the Olytics Behavior you want to modify, and then click, Manage Page Type Exclusions.

In the modal that appears, you will then be able to select the Page Types you want to exclude. The Page Type for each page will be populated from the “og: type” metadata present on the page and the options are Article, Profile, Book, and Video. Once you’ve made your selection, click Apply and pages with this type will no longer be pulled into the Content Recommendations program.

Excluding Specific Pages

To exclude specific pages from the Content Recommendations program, select the Olytics Behavior that want to modify the exclusions for and click, Manage Page Exclusions. A modal will appear where you can search for a specific page using exact match or wildcards (*) and then add the desired pages to the list of page exclusions.

When you’ve finished adding pages, click Apply, and those pages will no longer be recommended or included in the Content Recommendations program.

Excluding by URL String

To exclude many pages by a string contained within the url, this option can be used. Simply list the strings that, if found within the url, should exclude that url from recommendation. These strings should be comma delimitted.

Creating your Content Recommendations Personalization

To display the Content Recommendations on your site, you will need to create a Content Recommendations Personalization.

  1. Navigate to Personalization and Create New.
  2. Select the Personalization Type you’d like to use to display the Recommendations. Most users will select Standard, to display a pop up, or Inline, to display the recommendations in the context of each page.
  3. Expand the section labelled Content Recommendation Settings and select “This is a Content Recommendation Message”.
  4. Click “Generate Content Recommendation HTML”
  5. A modal will open where the user can select: the Behavior whose recommendations they’d like to display, the number of recommendations they’d like to display, the header text, and the header color. When finished, click Generate HTML.
  6. The HTML will populate in the content with the recommendation values populated as merge variables. The HTML can be changed like with any other personalization but be aware that the merge variables are what will be used to populate the links and link name. If these are removed, they will not work as expected.
  7. Click Save and Test on My Site to view how the personalization will show on your page with the default recommendations populated.
  8. Activate.

After activating your Personalization, note that if the Content Recommendations program has not yet run, the personalization will not display.

Using Content Recommendations in Email Builder

To add recommended content to an Email Builder deployment, a user and an organization must have access to Content Recommendations in Email Builder. Once the user has access, when a deployment is created or edited, a new option will be available in the Additional Options and Information section of the Deployment Summary pleat called, “Use Content Recommendations”.

After “Use Content Recommendations” is selected in a deployment:

  1. Select which Content Recommendations Behavior should be used to populate the recommendations.
  2. Select whether images will be used in the email for each recommended link. Selecting this option will ensure that links without images will not be recommended.
  3. Select the Number of Recommendations you will use in the email.

When the Content Recommendations settings are completed and the Deployment Summary pleat is saved, Content Recommendation merge variables can now be inserted into the message content.

In the WYSIWYG editor, several merge variables for each link can be inserted via the Insert Content Recommendations tool.

  • cr_X_linkurl : this will be populated with the url for the recommended page.
  • cr_X_linkname : this will be populated with the name of the recommended page. This value will equal the title value in the page’s metadata.
  • cr_X_imgurl: this will be populated with the image for the recommended page. This value will equal the og:image value in the page’s metadata.
  • cr_X_description : this will be populated with the description of the recommended page. This value will equal the description value in the page’s metadata.

In the Designer, these merge variables will available to insert into content via the Merge Tags selector.

Note: the Image content element in the Designer is unable to render merge variables so values are cleared. Users can populate the content in the designer using a placeholder image and edit the html to contain the correct merge variables.

After Content Recommendations merge variables are entered into the content, users can set the default value for each merge variable. These values will only be used for recipients who don’t have individual recommended links AND where there are no available top link to recommend. These default values can be set at the deployment level or in Deployment Defaults

Frequently Asked Questions

What is Collaborative Filtering?

Collaborative Filtering is a technique for predicting the interests of an audience in order to make a recommendation. Generally, the audience is placed into cohorts based on previous similar behaviors, in this case, page visits, and then a new recommendation is made to members in those groups for a page that has not yet been visited. To put it very simply, if Audience Member A visited Page 1 and Audience Member B visited Page 1 and Page 2, we might recommend Page 2 to Audience Member A. This technique is frequently used across many different web applications, like Netflix’s Top Picks recommendations.

Will recommendations show in my Personalization or Email for a new page visitor?

Yes! If a visitor is new to your site, recommendations will be shown using the top recommended pages across all cohorts.

Will my recommendations exist for Anonymous visitors?

Yes, if an Anonymous visitor has visited your page before, they will be given recommendations based on their previous behaviors and the cohort they are in. If they have not, they will be shown the most recommended pages across all cohorts.

Table of Contents
Scroll to Top