Recommendation Engines are systems that typically use Machine Learning to predict which movie or video a particular user (or cohort) is likely to enjoy watching, based on their past choices, preferences, and the content provider’s catalog.
Recommendation Engines are important for OTT platforms as they are a tool to help users navigate through a movie catalog efficiently. With the help of Machine Learning, platforms can build a persona for every user based on their interaction with the service, their choice of movies, and extensive movie metadata. Learn more about recommendation engines here.
In this article, we talk to Haris, the VP of Product and Data Science at Discovery Inc., to get an idea about their approach to and the use of recommendation engines in their video infrastructure.
Haris Husain [LinkedIn] is currently a VP at Discovery, where he’s focused on initiatives that help customers stay engaged with Discovery products. Previously, he led Building Technology at WeWork, where he developed new services that helped WeWork customers get the most out of their spaces. Before this, he led various programs at Amazon Search that helped customers find the best products for their needs. He spent much of his early career at Thomson Reuters, where, as an Engineering and later Product leader, he built products that helped financial industry clients work more efficiently. He enjoys building high-functioning teams and products that delight consumers and transform businesses.
With this introduction, let’s get started. Thank you Husain for agreeing to this interview. We are very keen to get your thoughts on recommendation engines, personalization, and the approaches to obtain, clean, and use data from different sources for your products.
Krishna: What is your approach to recommendation and personalization engines? Buy or DIY? Why?
Haris: We have started with 3rd party engines in the short term, and our long-term approach is to operate a blend of our own algorithms and 3rd party engines.
As our product features evolve, we will need personalization capabilities specifically tuned for our needs. We expect at this point it will become cost-effective for us to utilize our own models in addition to those provided by 3rd parties.
Krishna: What are your different sources of information, and where do you get them from?
Haris: Our basics are the typical set of catalog, customer information, and interactions between the customers and the catalog. Beyond these, we use a variety of internal and external sources for extra information about our customers and content that helps our algorithms become better at generating recommendations.
Krishna: How do you handle metadata creation on your end? Did you have to make changes in your metadata creation workflow due to the requirements from your personalization engine?
Haris: We use a combination of metadata generated at source by content producers, enriched by taxonomists, and enhanced by machine learning algorithms. To support personalization, we had to build data pipelines and workflows to enable this flow to happen efficiently and scalably.
Krishna: Can you share any figures from before & after using a personalization engine vis-a-vis engagement?
Haris: We cannot share exact figures; however, we can attest that personalized recommendations are among the highest drivers of engagement in our apps.
Krishna: What are the different ways in which you use a recommendation engine? Is it only for suggesting movies or any other use cases?
Haris: We have started with recommendations; however, we have other uses in the pipeline that will personalize more aspects of the customer experience. Stay tuned!
Krishna: As a final note to this interview, Husain, would you like to say anything to OTTVerse’s readers?
“Our customers tell us their interests and tastes through their actions on our apps, and using machine learning we help them discover shows that would appeal to them, and discover other content that they could develop a passion for. This helps them stay delighted and engaged with our products.”
Thank you, Husain, for this interview and for giving us an insight into the use of recommendation engines and personalization at Discovery Inc. We wish you all the best and look forward to more innovations from your team!
OTTVerse would also like to thank Satish Annapureddy for helping facilitate this interview.
Krishna Rao Vijayanagar
I’m Dr. Krishna Rao Vijayanagar, founder of OTTVerse. I have a Ph.D. in Video Compression from the Illinois Institute of Technology, and I have worked on Video Compression (AVC, HEVC, MultiView Plus Depth), ABR streaming, and Video Analytics (QoE, Content & Audience, and Ad) for several years.
I hope to use my experience and love for video streaming to bring you information and insights into the OTT universe.
Be the first to comment