Personalizing Cable Media
with Operational Intelligence

ScaleOut’s operational intelligence can track viewer behavior in real time and enable personalized recommendations—boosting viewer satisfaction and profits for media providers.

Mining Set-Top Box Data for Operational Intelligence

Millions of set-top boxes streaming data 24/7

Cable TV and “over the top” media companies continuously collect data from set-top boxes or viewing apps, which can be used as the source of operational intelligence. By streaming this data into an in-memory computing engine and enriching it with programming and historical information, these companies can generate real-time, personalized recommendations and highly targeted advertising.

Real-Time Content Recommendations

Providing personalized content recommendations when a viewer interacts with a media device increases both engagement and loyalty.

Programmatic Advertising

Operational intelligence implemented by immediately analyzing rapidly changing data enables media providers to deliver highly relevant ads in the moment.

Meeting the Need for Operational Intelligence

The cable TV industry faces ever growing competition from “over the top” (OTT) content providers like Netflix and Hulu, who are gaining market share and competing with each other, while advertising dollars are being split over these competing digital venues. This landscape has challenged media providers to find ways they can leverage their unique data to personalize the viewing experience and create relevant advertising for their subscribers.

Operational intelligence enabled by in-memory computing meets this need. Running on an elastic cluster of commodity servers, ScaleOut’s in-memory technology can ingest streams of real-time data from set-top boxes, correlate it with historical and account data, analyze it in-memory, and deliver personalized content recommendations and advertising to subscribers with sub-second latency.

The Opportunity for Personalized Advertising

Although advertisers now bid in real-time on web sites, media companies typically target ads to groups of similar people instead of individuals. By analyzing unique viewing profiles and history in real-time, they can increase the relevancy and effectiveness of their ads.

Targeting Ads with Operational Intelligence

Operational intelligence can combine set-top box data with historical patterns to form a real-time model of viewer behavior which is used by bidding systems to evaluate and select ads — all within milliseconds.

The Power of In-Memory Computing for Operational Intelligence

To deliver real-time, personalized recommendations and advertising for cable TV and OTT media providers, an in-memory computing system must be capable of processing huge amounts of data — fast. Because of its extremely low latency and scalable performance, ScaleOut’s in-memory data grid (IMDG) with integrated, in-memory computing provides an ideal platform for ingesting and analyzing this rapidly changing data. As shown in the following diagram, the IMDG captures real-time state changes from the set-top boxes and maintains an in-memory representation of viewer behavior. It enriches this data with historical information stored on disk, analyzes it within the in-memory computing engine, and immediately feeds the results to an external recommendation engine and/or advertising platform.

Analyzing Set-Top Box Data

In-Memory Computing System Used to Analyze Real-Time Data from Set-Top Boxes

To demonstrate that ScaleOut’s in-memory computing technology can deliver the performance required by a typical cable TV provider, a test was conducted to track, correlate, and analyze streams of data from 10M simulated cable TV set-top boxes generating 25K events per second. ScaleOut StateServer Pro performed these tasks with sub-second latency using a cluster of 12 commodity servers hosted on Amazon EC2’s cloud. It simultaneously analyzed data from all 10M set-top boxes every 10 seconds to display patterns and trends (such as the most popular shows by zip code). The analyzed data provides the basis for modeling viewer behavior in real time, making program recommendations, and selecting ads to display.

Try ScaleOut for free

Use the power of in-memory computing in minutes on Windows or Linux.

Try for Free

Not ready to download?