Use Cases and Tasks

Rumo sets the path from personalization to content suggestion.

It does not need to be aware of the content type, or the delivery format (on-demand, or live) to make content recommendations. However, the source you upload to Rumo serves as the basis of all recommendations.

We consider five dimensions to personalization (the 5 Ws):

  1. Who to personalize – to individual users or user segments using our technology, through implicitly deriving personalization from the user’s tastes, or explicitly integrating with your feedback systems, such as like/dislike or star-rating systems.
  2. What to personalize – the content itself, the channel, or the interface through finding content the user enjoys.
  3. When to personalize – in real-time based on past interactions, or taking into account the time of the day, the day of the week, or others.
  4. How to personalize – through semantic data related to the content.
  5. Why personalize – to provide accurate, trustworthy, and transparent content recommendations based on content knowledge.

Let us consider a general overview of the main use cases you may find while using Rumo.

Use Cases Description Tasks
Authentication Rumo requires an account to submit queries, and keys to authenticate requests.

To authenticate with Rumo, name the x-api-key key in the header of your requests after the API key we send you during your registration process.
1. Join Rumo
2. Check the terms and conditions
3. Wait for us to send you the information you’ll need to proceed
4. Start using Rumo by uploading your content and submitting queries
Similar content-to-content (C2C) recommendations We work the suggestions at the content-level, based on the previously consumed content.

We use the keywords available to find similarities between content pieces. The better the content database is structured (through the keywords, categories, and weights that you specify), the better the recommendations that Rumo provide.
1. Submit content
2. Retrieve similar content
Personalized user-to-content (U2C) recommendations We define the semantic user fingerprint, which is the user’s taste profile, based on the content with which the user has interacted. We then refine the content recommendations using the defined fingerprint. We can also use the user’s feedback by fetching the data collected through your rating or feedback systems. 1. Submit content
2. Submit user events
3. Retrieve personalized recommendation
Manage content You will progressively have access to more data points related to your end-users and your account activities. 1. Submit content
2. Retrieve content
3. Retrieve content piece
4. Update and deleting content