How Netflix Used Big Data To Achieve Phenomenal Growth

Source: Pixabay

Netflix’s phenomenal worldwide growth is a significant factor in the company’s continuing success. As of 2017, it was operational in more than 190 countries, and of its 130 million or so subscribers, almost 73 million reside in jurisdictions outside the USA. And by Q2 2018, Netflix international streaming revenues had overtaken its US streaming revenues. Bearing in mind that Netflix was a US-only business until 2010 and had a presence in just 50 countries in 2015, this constitutes an outstanding business achievement. It is therefore no surprise that many large industries, including online casinos and e-commerce sites, now follow the Netflix model in their use of data analysis strategies.

The Netflix globalisation strategy has managed to overcome some tough problems: Different regions demanded different content deals, diverse legislative barriers often had to be approached country-by-country, many subscriber blocks were non-English speakers and insisted upon local content, and many would-be subscribers had only ever experienced free content and were very reluctant to sign up to paid-for streamed content.

Despite all these problems and more, Netflix has managed to grow consistently and exponentially and is now a global entity with more users of paid-for content than the total market share of all its streaming rivals combined.

Big data tools

One significant feature of the Netflix story has been the company’s continuing sophisticated use of data collection and analysis tools to inform its business decisions. These have been deployed to gather, map and interpret findings based on a wealth of data Netflix has collected about the preferences of a host of different user groups. Another by product of this approach has been that whilst the Netflix analysis budget has gone up, its marketing budget has been significantly reduced. One result of this change is that Netflix is therefore marketing to targeted groups and populations who have an existing interest, or at least a strong potential interest, in Netflix products. This means that the company rarely waste money on any kind of purely ‘scatter gun’ speculative marketing to users who may have no direct interest in making any kind of purchase decision in their favour.

The recommendation algorithm

The monitoring tools Netflix uses are always gathering information and continuously testing out scenarios based on their in-depth analysis of the Netflix data flow. So each Netflix subscriber click on play, pause and stop is recorded and scrutinised in detail.

It is believed Netflix runs around 250 A/B tests per annum. These tests involve about 100,000 users, plus another 100,000 selected as a control group. They are streamed the same content but with slight variances in the format. Such tweaks might be in the design or look of the streamed content, and Netflix are interested in assessing how such adjustments are received by their user audience.

Another area of appraisal is the use of landing cards – the display images Netflix users will find as they browse through lists of film titles available on Netflix. The company is once again seeking to evaluate the effect of slight format changes on user behaviour. It seems Netflix are also primed to introduce the same system with its autoplay trailers as a means of identifying the most popular options.

Statistics suggest that an average viewer will browse around 50 titles before they are ready to select the next film they want to stream. And it appears that the majority of Netflix analysis is presently being deployed to fine tune that list of options to match the viewer’s likely preferences. Each list is user-specific, and is primarily based on previous watching history but continuously informed by any other data trends which have been noted.

For instance, Netflix will be aware of what time of day you tend to watch, and that information will have some influence on the choice of films the company then choose to offer – though Netflix are a little reluctant to detail what tweaks they are likely to make, and why.

Case history: House of Cards

Source: Pixabay

Here, Netflix noted a great many of its subscribers tended to stream ‘The Social Network’, directed by David Fincher, right through from beginning to end. They also observed that Kevin Spacey films had a consistent appeal for the Netflix audience. Further Netflix analytics indicated that the UK version of ‘House of Cards’ was a success and revealed that British viewers also sought out other movies directed by David Fincher or featuring Kevin Spacey.

Based on this data assessment, the company predicted that ‘House of Cards’ would also be a success in the USA and decided to invest $100 million accordingly. This smart decision was a resounding success, gaining Netflix 2 million extra US subscribers plus 1 million more worldwide, all within the first three months of the launch. As a result, Netflix had virtually covered its ‘House of Cards’ investment within the first quarter.

Investing in big data analysis has enabled Netflix to target and customise its streamed offerings to such an extent that, as with ‘House of Cards’, the figures can also inform future decisions about investing in movies and TV series. Effectively limiting, or even removing, such risks has thus been a major factor in the phenomenal growth of this ambitious company.