Announcing the Tweek Entertainment API
At Tweek we’ve spent a long time iterating on how to provide our users with the most relevant entertainment recommendations. We build our own app aggregating 20+ Video on Demand sources into a personalised content stream and integrated different recommender types along the way to understand consumer needs and behaviour. Today, we’re releasing a building block of our technology, the Tweek Entertainment API.
The Entertainment API helps to personalise client content offerings towards the consumer, enabling personalisation and recommendation features. Features which are becoming increasingly important as they drive revenue for companies like Google, Facebook, Amazon, Netflix, and LinkedIn. Those companies are reporting increased revenue and engagement with machine learning products for years. Offering personalised and tailored, digital services to their existing audience is at the core of their services to keep their consumers engaged. Our API enables media companies to include similar features without building the core technology themselves.
Also, we are happy and proud to announce our first Entertainment API client and successful integration of our technology: Together with MyVideo (the web/ mobile subsidiary of Pro7Sat1 serving 10m+ monthly active users) we increased music video views for logged in users by 143%. The service is monetised through video ads so video views are the most important KPI influencing revenue results, directly. Think about: What would this increase of view count mean for your revenue?!
For the consumer, our personalisation offers a tremendous increase of user experience. The music video catalogue of MyVideo exceeds 100k items. It is impossible to navigate and users are often in a non-search mode but just want to lean back and have some music video playing which they like without going lean forward into search, thinking anything specific.
Our new API provides the knowledge about a user and her/his Entertainment Graph (this is how we call it) including film, tv shows and also music to our clients. It is capable to read existing, entertainment related data signals of the user from the social web, connect them to our proprietary ontology and — if needed — calculate recommendations on top. Those can be matched against your content catalogue and delivered through a simple and easy to integrate API.
Tweek has aggregated, normalised and enhanced entertainment data since 2011. By doing this we were able to harvest a vast amount of signals which helped us understand patterns and ultimately predict behaviour.
We accelerated our recommender technology development with a cross-functional research team in cooperation with German Research Center for Artificial Intelligence during the last two years. We created collaborative filtering, meta data based and semantic recommender entities. The cooperation helped us to scale vertically across entertainment fields in parallel: While our first focus was about TV, we added movies and TV shows, soon. Now, we are very excited to announce that we can connect music related data points to our taxonomy, too. Doubling the number of video views is a great proof!
We have received many requests from partners and developers to open our technology. Today, we’re happy to publish this first step: You can test our music API on top of the MyVideo music catalogue via demo.tweek.tv. Login with your Facebook account and receive personalised recommendations for MyVideo music videos within seconds. As a next step, drop us an email and we will provide an access token to our API.
We are excited to see the creative and new projects that will come from getting our API into the hands of developers. We are working on making it easier to access and wider in content coverage. It you want to stay up to date on our technology evolution, sign up for our company newsletter.
Please find today’s press release for download as a zip file here (German language).
Originally published at medium.com on September 9, 2015.
