Content Optimization™

The selection of content involves a strongly subjective component (i.e. Art versus Science).

Achieving the optimal mix of content is one of the most important tasks to ensure long-term ratings success.

It is difficult to consider multiple variables (+10) to explain content success using experience alone.

It is possible to predict how content will perform given a set of content attributes (i.e. High and low rating).

​

Key Objectives:

Provide additional inputs to the content selection and programming teams on the effect of key attributes on content performance through 2 mathematical models.

• Explanatory: Identify the attributes that historically explain the rating.

• Predictive: Given certain attributes, predict how certain content will perform.

​

By teaching the “Machine” how to identify the most important content attributes that maximize Rating (or Share) for a particular time, day or feed.

What is Metadata?

All the variables that could explain the KPI we want to maximize. For example:

Objective Metadata​

​

Production House

Number of episodes

Cast, ethnicity

Accent

Screenwriter

Broadcast time

Premiere date

Subjective Metadata

​

Subjective opinions

Romantic, Dramatic, Sexual, Religious.

Analytic strategy for explanatory model (example Decision Tree)

We seek to determine:

​

•The most important characteristics of a production in determining the level of success.

​

•The trail of attributes that lead a production to be {good, average or bad}.

Analytic strategy for predictive model

•We seek to determine if the proposed content will reflect a high or low rating through a classification model for a feed.

​

•Likelihood

    - High

    - Low