Art of Analyctics: Hollywood Nights

Learn Data Science
Teradata Employee

 

Capture.PNG

 

 

 

The Insights

What could movie or TV watching teach us? Turns out, a lot! Entertainment companies have placed exclusive focus, especially in an era of intense competitive pressure, on delivering unique customer experiences that are highly personalized to the idiosyncratic and shifting tastes of the viewing public. Gone are the days of a two sizes fits all entertainment packages where the onus is on the customer to cherry pick content from a static set of options. This leads to a widely felt dissonance in that viewing preferences are vastly unmatched by the actual entertainment offerings. When we add to this the diversity of how entertainment is delivered — through set top boxes, streaming platforms, digital applications, theaters — the issue becomes even more complex and urgent to address.

 

Each circle represents the viewer size and prominence of a specific program (or movies) or closely related set of programs (e.g., a series) over time. These circles are grouped into logical clusters (the spoke-like outgrowth that comprise of several closely linked circles in a sac). Some examples of programs that have resulted in the large circle include "Hunger Games" and "Avengers". Some of these circles are much older TV serials but still remain popular draws among certain segments of the viewing public. Furthermore, each circle shows a relationship with other circles sometimes in the same genre (sac) and sometimes by spanning entertainment genres. The thickness of the lines, connecting the circles indicates the number of times these programs have been seen together. Clusters can thus consist of a wide number of content pieces. Over time, these logical groupings of entertainment clusters and their relationships can be used to create highly personalized packages based on initial viewing preferences that are expressed by the subscriber to services (e.g., Cable TV). 

 

The Analytics

This visualization was created using Teradata Aster Analytics. We used detailed program viewership data (e.g., subscriber ID, programs watched, duration, repeat viewerships, program categories, geography, subscription duration, program release date) that were loaded into Aster Analytics for analysis. Several datasets were included to align video content consumption with actual customer households resulting in a massive dataset of more than 600 million rows of video content transactions on a catalogue of several thousand titles. Aster's native data prep and graph algorithms (e.g., modularity, degree) were used to determine logical program groups and the relationships within and between the groups. A visualization of the entertainment clusters was then rendered using native visualization capabilities on Teradata's AppCenter framework.

 

The Benefits

There are at least three benefits that are realized through this advanced analytics implementation:

 

First, entertainment companies can gain new and timely insights into the relevance of their broad content portfolios vis-á-vis their customers that can be segmented across many demographic and socio-economic groups. What this engenders is an ability to provide entertainment tiers that fit the demand profile and price elasticities of customers.

 

Second, customers are now more satisfied and can get creatively packaged content that not only fit their current preferences but also expose them to other content types that they could grow to appreciate.

 

Third, content providers can now expect more strategic support (e.g., royalty payments, production support) from their buyers given that a clear data driven view into entertainment content clusters has been established. They can now confidently work on generating viewing material that they know will be surefire hits with the viewing public.