JW Recommendations

To access these features, Upgrade Now
Optimizing Recommendations

JW Recommendations integrates with each part of JW Player's product to provide a powerful, real-time & automated content discovery solution built from our network of viewer and content data. 

JW Recommendations' Solutions:

  • Our Similar Video Playlist finds the best videos to complement what viewers are watching fro viewer engagement, machine learning semantic similarity, and real-time video trends. This powers our in-player recommendations UI (overlay & shelf) and leverages viewing behavior in the broader JW Network. It is available via our v2 Delivery API.
  • Our Video Search Playlist matches a keyword search against your property's content library and presents the best results. Search cannot be embedded in a play at this time however it is available via the v2 Delivery API.
Note that Search and Similar Playlists require a video or keyword to produce a playlist of relevant content. See below and our API documentation for more details.

What are benefits of JW Recommendations?

JW Recommendations can be easily implemented across any video publishing platform, is delivered through a beautiful & responsive in-player interface, and features great reporting & analytics.
  • Easy implementation with the dashboard player builder or API (JSON, RSS)
  • Full support for Contextual Page Widgets, OTT, Showcase, and Mobile SDKs
  • Measurable viewer engagement and overall impressions via Segmentation
  • Editorial control with Promote & include/exclude tag filtration rules
  • Lowered human curation effort and better automatic content discovery & recirculation 
  • Cost savings over in-house or third party solutions

How do I use JW Recommendations?

Publishers using JW Recommendations can provide viewers with recommended content in several ways:

  • In-Player: If you are using single-line embeds (Cloud-Hosted Player + JW Platform content) or cloud hosted libraries, you can easily turn on the Discover Overlay (& starting in 8.2 the Shelf interface) to begin displaying recommended content in any JW8 player.
  • JSON / RSS: Use the JW Platform API to leverage recommended content in a mobile app, OTT app, or page widget OR to access Recommendations programatically.

Start Using JW Recommendations Right Now...

You can immediately use JW Recommendations when building a player if:

  • You're a customer who has Recommendations enabled. (If you don't and would like to, contact us!)
  • You embed JW8 players via the dashboard-provided single-line embed code or using a cloud hosted player with libraries!
  • You EITHER host and stream with JW Platform OR, if not, you register content for Media IDs
    • Tip: Add relevant titles and descriptions for the best results

How can I get the most out of JW Recommendations?

Follow our best practices for video player & webpage user experience: 
  • Optimize placement of the player & page widgets to maximize interaction (i.e. above the fold)
    • Be sure to use large video players where possible!
  • Keep page load time fast, scripting light & build responsively for mobile viewers. Consider using our Showcase suite of solutions to make this easy and low development effort.
  • Choose good thumbnail images, they often lead to a big increase in engagement.
  • Enter relevant title & description metadata, and if possible keep internal business language in custom parameters or CMS tags.
  • Use high quality content, and if you use tag filtration try to limit to very carefully defined exclude rules (tag rules will greatly affect the engine and should only be used when required!)
  • As your testing be aware that the player will filter already viewed videos and re-recommend them based on algorithmically defined rules so you might 'repeat' recommendations and that is expected.
Implement a comprehensive monetization strategy to meet increased demand:
  • Optimize Ad tag placement and tag waterfalling to ensure a good viewer experience.
  • Pay attention to Ad fill rates and demand to prevent lost monetization opportunities.
Be sure to use Recommended playlist content everywhere:
  • Turn on the Recommendations UI for all your players, even when embedding manual or dynamic playlists. (It gives viewers an extra chance to discover more content after your playlist or single video completes!)
  • Ensure that your Similar Video Playlists all have at least a 25 item count. 
  • Place any out-player page widgets where viewers can see them while watching the in-player video.
  • Build contextual players with keywords on article pages to highlight other content viewers might be interested in.
  • Swap out Dynamic and Manual Playlists with Trending Playlists where possible to lower the workload on editorial staff.
Measure performance and tune for your business and publishing strategies:
  • Our Analytics features like Segmentation and Labs can help you understand the performance of specific recommended playlists and content types.
  • Leverage our Promote feature to ensure specific videos are featured in the first or second positions by our recommendations engine and drive additional views.
  • Experiment with different types of recommendations for different pages, players, and apps.
  • Use our flexible system of tag based business rules to construct ideal recommendation feeds for your viewers by continually experimenting and keeping the strategies that work. Contact support@jwplayer.com for any questions or more help.

What else should I know?

There are some important things to keep in mind about the recommendation product:
  1. Sparse results could mean you have insufficient user traffic levels, sparse content metadata, or your content library (or tag filtered content pool) might be too small. JW Recommendations Engine can only recommend content well when these factors are fully considered. JW Player Support can help you troubleshoot issues and offer advice if you have sparse coverage.
  2. Similar Videos and Video Search Playlists require the JW Player (Automatic for Similar Video Playlist in-player if you've turned on recommendations from our player builder) or developer/publisher (using JW Platform API) to send a request to the JW Recommendations API with either a Media ID (Similar) or Keyword/Phrase (Search) for the result set/playlist to be built. See our API documentation for more details.

Did you find this article helpful?

Please log in to rate this article.