Entitlements Caching

Approach overview

Broadcasters approach entitlements in different ways:

  • Some of them choose to send API requests each time a customer attempts to access content.
  • Others prefer caching entitlements. This may be caused by a few reasons, e.g. they might want to have an additional security layer or to have more control over entitlements. Another reason might be technical dependencies - if they have their own customer authorization mechanism and need to have their own token with entitlements.

Both approaches are supported by Cleeng, depending on broadcasters’ preferences.

Cleeng entitlement API is optimized and prepared for a big number of requests, however caching entitlements can help to avoid a flood of multiple requests from the same customer (end user) during popular video streams (e.g. because of multiple page refreshes).

πŸ“˜

Important

The way how the Cleeng API is implemented and your caching strategy can affect the numbers that are reported for engagement or churn statistics.

Caching entitlements vs churn prediction

It is worth noting that in order for the machine learning model to have enough data to predict churn the following conditions must be met:

  • 100 new subscribers in the last 4 weeks
  • 100 churned subscribers in the last 4 weeks

In order to use the churn prediction model and engagement dashboard, your Cleeng integration must generate engagement data through the Cleeng entitlement engine.

In technical terms, we understand engagement as a successful entitlement check for a customer. A successful check occurs when:

As a result, the exact engagement numbers will depend on how the Cleeng API is implemented, e.g. it can be affected by your entitlement caching strategy.

When, as a broadcaster, you cache entitlements on your side, and Cleeng API is not called, then Cleeng lacks information that is necessary for the correct engagement and churn calculation.

To solve this problem we recommend that you should send a request to one of API endpoints (enumerated below) that are counted as engagement at least once every 24 hours when a customer attempts to get access to content (e.g. clicks on a "Play" button) so that the Cleeng algorithm has the right data to predict churn.

Endpoints counted as engagement: