Facebook: Big Data company based on SNS

The big data age has come and big data is becoming mainstream. According to IDC, the global market for large data technology and services will grow from $3.2 billion trillion in 2010 to $16.9 billion in 2015, with a composite annual growth rate of 40% per cent, about 7 times times the overall information and communication technology market. The influence of large data on the development of enterprises can not be neglected.

2012 was a year of big data outbreaks, with many tech giants flocking to big data areas, and start-ups based on big data, in which Facebook has been making a big deal this year, seizing opportunities and becoming one of the most deserving leaders in the data world.

Facebook's Big Data strategy

Facebook has always been a great growth story from social giants to big data companies, but the root cause of its success is not social, but the massive data and large data strategies it generates from its social networks.

May 18, 2012, Facebook listed on the Nasdaq, IPO priced at 38 U.S. dollars, financing on the scale of 16 billion U.S. dollars, and at the price of the price of Facebook, the value of 104 billion U.S. dollars, as the United States has been listed in the market capitalisation of the largest companies. Facebook's IPO represents the third wave of the global Internet business model, and its IPO leads the internet into the big data age, according to a recommendation issued by the Research Department of CITIC Securities (600030, shares).

Data collection, data analysis, and data applications are the troika that makes up Facebook's big data strategy.

Data collection is Facebook's forte, and since its inception, Facebook has been consciously collecting user data. Facebook's large data product, timeline, released late last year, allows users to record their life stories on the Timeline page, and to extend the range of data collected from Facebook to historical data. In 2012, Facebook also brought a lot of segmentation data to its acquisitions by companies such as photo-sharing app Instagram.

In an era of almost universal social networking, social-networking Facebook has become a well-deserved data-gathering giant, with "massive + unstructured/complex types" of data generated from these social networks, indirectly driving the development of large data industries.

But just collecting data does not produce value, mining the commercial value behind these massive data becomes an integral part of the big data industry chain.

As some academics say: "Facebook's efforts over the years have created links and ties to more than 1 billion digital migrants, and the world's borders are still expanding, and the next step is to consider how to make the massive data generated by the relationship more valuable." ”

After data collection, Facebook needs to classify and structure the messy data, then interpret it and analyze it to get the specific information of the data users and lay a good foundation for large data applications.

"What we do boils down to just one thing, the big data processing," says Jay Parikh, vice president of infrastructure technology at Facebook. "The big numbers are leveraging your insights, discovering the potential value behind the data, and using it to create benefits for your business." The simple truth is that if you don't make good use of the data that you collect and store in your business, then you have just a bunch of data, and what we're more and more interested in doing is studying how to use the collected data to do something more worthwhile. ”

Data applications are not fully stereotyped in Facebook's big data strategy, focusing on three levels of advertising marketing, product services and user management.

Through the early data collection and analysis, Facebook can understand the needs of users, timely adjustment of product design and services to meet the needs of different users.

At the same time, the data on the advertising is also very useful, Facebook can work out a targeted advertising program, and targeted to the users of the marketing campaign. Advertising marketing, one of Facebook's main profit models, has been one of the main developments in Facebook's big data strategy, and big data can help Facebook get more advertisers and create more value for the data.

As Facebook reported in its third-quarter earnings in 2012, revenue from the advertising business accounted for 86% of the $1.262 billion trillion in the third quarter, up 36% from a year earlier.

Facebook's core data assets

In the eyes of investors, they prefer to think of Facebook as a big data concept relative to the SNS concept, because Facebook's core competency lies in its core data assets, And Facebook's Oregon State Prineville data Center is a great indication of its own massive data.

Facebook disclosed a set of data this August, with 2.5 billion messages per day, 500+ TB of data, 2.7 billion clicks on the Like button, 300 million users uploading photos, and about 105TB of data scanned every half hour via Hive. And Facebook has the world's largest single Hadoop system, with more than petabytes of data stored in each individual Hadoop disk cluster, even at today's fast pace, and perhaps soon, the petabytes of disk clusters are no longer surprising.

At present, Facebook has more than 1 billion users worldwide, 44% of the world's Internet users use Facebook, mobile users up to 600 million, the United States, Brazil, India is used in the first three countries, a total of 140.3 billion friends connected, photos uploaded to 219 billion, 62.6 million music was played, Up to 22 billion times.

Facebook's social networks generate huge amounts of data every second, such as text, pictures, music, video, geo-information, web links, and these data are updated, aggregated, and not crawled by search engines, making Facebook's core data asset.

Even though Facebook was not entirely aware of the usefulness of the data collected at the outset, they simply "wanted to know who visited the site, their dynamics, and what section of their site they were on, so it was all interesting to us." Jay Parikh said. But as the value behind the data is being discovered, Facebook is increasingly focusing on collecting web data and laying down its own data base.

While simply collecting data without analyzing it, it cannot reflect the real value of the data, nor can it be called large data. "The point of the big data is that you really have an inherent insight into your business. If you can't make good use of the data you collect, you're just a bunch of data, not big data. Jay Parikh said.

For these massive amounts of data, Facebook captures the real value behind them, not just using it rudely, but by analyzing and applying the data to really achieve big data. From the raw material producers to the raw material processors and users, the use of data is not only a step.

Facebook, as a social media, has an innate advantage in data collection, with more than 1 billion of users bringing a huge amount of data to Facebook, an invaluable asset in an era when the big data is about to explode, and Facebook is seizing the opportunity In the 2012, a thick-fat, the digital wealth into a real wealth.