How Netflix Uses Image Analysis To Create Visual Content

Producing high-quality original content is expensive, so Netflix analyzes its vast collection of data to increase the odds of creating content that appeals to its subscribers. With subscribers presented with nearly unlimited options, why leave such a potentially critical aspect completely to chance? After all, Netflix possesses the data to make the most informed business decision possible.

In a previous Blog Post we talked about how Netflix is using massive amounts of data to make intelligent decisions on creating original content.

They also perform data analysis on images to gain competitive advantage.

Questions Netflix answers with Visual Image Analysis include:

– Are certain customers trending toward specific types of covers? If so, should personalized recommendations automatically change?

– Which title colors appeal to which customers?

– Is there an ideal cover for an original series? Or should different colors be used for different audiences?

Look at the covers of their original series House of Cards, and the 2010 version of Macbeth that ran on the PBS series Great Performances.

They are very similar in terms of color balance, contrast, and visual field. They both display older white men with blood on their hands(Kevin Spacey and Patrick Stewart). Both images have primarily black backgrounds, and use a very similar color pallet.

The covers of the two shows are much more similar than dissimilar. However, subtle differences do exist between the two images, and Netflix can precisely measure those differences. The next image details how Netflix visually interprets the data from those images.

Netflix measures if the similarities and differences have any quantifiable impact on subscriber viewing habits, recommendations, and ratings. Management teams carefully review this data when selecting the covers for their original content series.

What is next? Comparing the visual contrast and hues of images isn’t a one-time experi­ment for Netflix, it is a regular component of their strategy. Netflix recognizes that there is tremendous potential value in these discoveries. To that end, the company has created data analysis tools to unlock that value.

At the Hadoop Summit in 2013, Netflix employees Magnusson and Smith talked about how data on titles, colors, and covers helps Netflix in many ways. For one, analyz­ing colors allows the company to measure the distance between customers. It can also determine, the “average color of titles for each customer in a 216-degree vector over the last N days.”

A great example of a company transforming data into intelligent, game-changing results.

REFERENCES: http://techblog.netflix.com/

How Netflix Uses Big Data To Create Content

Netflix collects an enormous amount of data on its users and their actions, and uses that data in ways never before used in the entertainment industry to decide what content to produce. The numbers are staggering: Netflix is the world’s leading Internet television network with over 57 million members in nearly 50 countries enjoying more than two billion hours of TV shows and movies per month.

Netflix collects an enormous amount of data on its users and their actions, and uses that data in ways never before used in the entertainment industry to decide what content to produce.

The numbers are staggering: Netflix is the world’s leading Internet television network with over 57 million members in nearly 50 countries enjoying more than two billion hours of TV shows and movies per month.

Netflix looks at data generated from over 40 million “plays” a day. It keeps a record of every time subscribers pause the action, rewind, or fast-forward. They track how many subscribers abandon a show entirely after watching for a few minutes. In addition they track the time of day when shows are watched, and on what devices they are watched on.

Netflix user account data provides verified personal information (sex, age, location), as well as preferences (viewing history, bookmarks, Facebook likes).

Having this wealth of detailed knowledge of Netflix subscribers, what they watch, and how they watch it allowed the company take the next step of creating content, but do it using intelligent data driven decisions.

The data collected by Netflix indicated there was a strong interest for a remake of the BBC miniseries House of Cards. These viewers also enjoyed movies by Kevin Spacey, and those directed by David Fincher.

Netflix determined that the overlap of these three areas would make House of Cards a successful entry into original programming.

The result: House of Cards is the most streamed piece of content in the United States and 40 other countries, according to Netflix. It is rated 9.1/10 from 180,816 users at IMDB.

In any business, the ability to accurately predict what products or services will succeed is of paramount importance and value.

What is next? Correlating the raw numbers of Kevin Spacey fans who also like David Fincher and have a fondness for British political dramas is just the beginning. Netflix’s deep dimensions of data about what you are watching can be used to judge specific aspects of content as well. Senior data scientist Mohammad Sabah reported at a Hadoop Summit conference in 2012 that Netflix was capturing specific screen shots to analyze in-the-moment viewing habits, and that the company was “looking to take into account other characteristics” as well. What could those characteristics be? GigaOm’s report of the Sabah presentation speculated that “it could make a lot of sense to consider things such as volume, colors and scenery that might give valuable signals about what viewers like.”

In the next post I will explore how Netflix uses Data Visualization to make Intelligent Design Decisions.

REFERENCES: http://techblog.netflix.com/