Insight Platform

Insight Platform

End-to-end collection, cleansing, warehousing and analysis of viewing data

The Insight Platform is the beating heart of the Genius Digital system. It:

  • Collects return path viewing data from a customer’s device
  • Cleanses it of errors and other commonly found issues in big data sets
  • Fuses it with non-viewing data

Data collection

The Insight Platform collects data from any devices that a customer watches a TV or video service on. This includes:

  • Tablets and smartphones, using our iOS and Android APIs and libraries
  • Browser-based devices, including PCs, TVs and newer set-top boxes with browsers, games consoles and media streaming devices such as Chromecast, using our JavaScript APIs and library
  • Set-top boxes: If a service provider’s middleware provider is already collecting viewing data, we can receive data from here. We can also use our MiriMON collection agent to collect data from an operator’s set-top box.  This allows service providers to collect real-time data

Data cleansing

Every dataset contains errors, duplicates and abnormalities. At the data cleansing stage, the Insight Platform removes these from the collected data and normalises data collected from different devices in a service provider’s network.

Some of the common issues that the Insights Platform fixes at the cleansing stage include:

  • Errors around timeshifting (data suggesting a programme has been viewed before transmission)
  • Duplicates, where a device watches the same channel at the same time more than once
  • Matching schedule and consumption information
  • Verifying that key identifiers match across all data sets
  • Looking for gaps and overlaps in schedule data
  • Genius Digital’s truncation model, which removes instances where a TV has been left on, but content is not being viewed, is also applied here.

Data fusion

At the data fusion stage, the Insight Platform matches other operator owned data-sets to the cleansed return-path viewing data. Once service providers have fused the return-path viewing data with other datasets, they can segment and cluster the viewing data much more effectively. The more datasets we are able to fuse, the greater the ability an operator has to target and segment subscriber base.

Typical datasets include:

  • Subscriber demographics and CRM information
  • Call centre records
  • Operational history
  • Customer segmentations
  • Billing information
  • Churn model information
  • Customer lifecycle information
  • Programme metadata
  • Package information

To learn more, download a product brief here: