In competitive markets, pay-TV operators continually strive to increase their market share - not to mention retain their existing customer base - by providing more and better content. But the downside is that the more channels and content assets there are, the more the consumer is required to monitor channels, find new programmes and decide what to watch.
TV is a recreational activity; consumers don't want to work too hard for it. Statistics show that, despite having hundreds of channels at their disposal, the average pay-TV user watches fewer than 10 channels on a regular basis. There are various reasons for that, one of the most important being that most users are unable to keep tabs on all the programming that is available. Invariably, users stick with what they know.
It's a dilemma faced by virtually every operator. Differentiation in the market is critical and content acquisition is expensive, but operators lack the means to make programme access easier for the consumer, draw the consumer's attention to relevant content and effectively monitor usage to ensure that the investment in additional channels and premium content is paying off.
Consumers often churn or cancel premium packages because they are unaware of the range and quality of the programming available from their pay-TV operator. Some are giving up their pay subscriptions altogether and migrating to digital free-to-air services, many of which offer dozens of channels.
Promotion without confusion
Customer retention and increased ARPU, primarily by means of bundle upgrades and diversified pay services, are critical for pay-TV operators. They urgently need new technologies to assist them in promoting their content and services to consumers, while, at the same time, reducing consumer frustration by cutting the time and effort required to find the content they want.
One such technological capability is recommendations, a tool that is well established on the Web but which is still in its infancy on TV.
Recommendations do exactly what the word implies. They recommend content to a user or groups of users, based on content that has already been viewed by the user or information about the user's interests. The viewing suggestions can be provided by the operator, other consumers or third parties and they can be displayed in a variety of ways. What all recommendation techniques have in common is that they are a means of up-selling additional content and services on the basis of some form of knowledge of the customer.
Online, recommendations have a proven track-record. Fully 35% of Amazon's product sales are the result of recommendations, which have shown themselves to be vastly more effective than any other means of promotion deployed by the company. On Google News, recommended items receive 38% more click-throughs than those which are not recommended.
The challenge of recommendations on TV
The challenge is to reproduce the same effectiveness on TV networks, the bulk of which are still one-way only and do not support the user interaction that is common on the Web.
In general, recommendations can be created manually by the TV operator or generated by algorithms, which compare and match the "DNA" (or metadata) of content and users. Otherwise, they can be based on ratings provided by users or third-party agencies.
NDS has developed a recommendations concept framework which is capable of supporting all the above techniques and is suitable for both unidirectional and bidirectional networks, according to NDS Product Manager Alon Weinberg. It is an end-to-end system, which is synchronised with the broadcast schedule and can be operated manually by the TV broadcaster or can integrate with a third-party recommendations engine.
Operators need technologies which assist them in promoting their content and services to consumers, while, at the same time, reducing consumer frustration
Crucially, it is built for the mass market. It is non-intrusive, contextual ("if you like this, you may be interested in …") and simple for the consumer to use. Users can be divided into sub-groups, which receive different recommendations, and usage of the recommendations can be monitored using the NDS Audience Measurement System.
The proof of concept (POC) for recommendations was developed by the NDS Marketing Technologies Group, a team dedicated to developing new functionality concepts and demonstrations. Designed for a one-way broadcast network and running on a STB with Fusion middleware, the POC comprises headend metadata components, an Operator Control Panel, a STB user interface and integration with a recommendation engine from ThinkAnalytics.
The ThinkAnalytics backend provides recommendations based on context (e.g. most recently viewed item) targeting to subscribers (two-way) or subscriber groups (broadcast) or individual preferences if available. The NDS headend synchronises the recommendations with broadcast and on-demand playout. Viewers see a real-time interactive recommendation application on the STB.
In the future, when IP and hybrid STBs are more widely deployed, the recommendations concept framework will support user profiles and ratings, recommendations according to profile, content push based on profiles and a comprehensive audience measurement solution.
|
