If you were asked to think of one company that is defined by its use of algorithms, you might name Google.

And for good reason: the search giant's algorithms are not only at the heart of its success, but for many, they're the source of constant hope and fear as changes to them can literally make or break businesses.

But another prominent name on the consumer internet might also be a viable contender for the title 'algorithm company': Netflix.

In a post on its blog, Netflix revealed details around the recommendation algorithms it uses. Note the plural 'algorithms'. Netflix personalization science and engineering staff members Xavier Amatriain and Justin Basilico explain:

Personalization starts on our homepage, which consists of groups of videos arranged in horizontal rows. Each row has a title that conveys the intended meaningful connection between the videos in that group. Most of our personalization is based on the way we select rows, how we determine what items to include in them, and in what order to place those items.

Take as a first example the Top 10 row: this is our best guess at the ten titles you are most likely to enjoy. Of course, when we say “you”, we really mean everyone in your household. It is important to keep in mind that Netflix’ personalization is intended to handle a household that is likely to have different people with different tastes. That is why when you see your Top10, you are likely to discover items for dad, mom, the kids, or the whole family. Even for a single person household we want to appeal to your range of interests and moods. To achieve this, in many parts of our system we are not only optimizing for accuracy, but also for diversity.

That's just the beginning. Netflix's algorithms also factor in awareness, freshness, similarity, and social connections, amongst other things.

Why the sophistication? Miraculously, Netflix says that "75% of what people watch is from some sort of recommendation." From this perspective, it's not a stretch to say that Netflix's business today is driven by the ability of its recommendation algorithms to make good recommendations. Which sort of explains why the company created a million-dollar challenge with the goal of improving its recommendations by what appeared to be a small margin.

While most companies aren't as big as Netflix, many businesses large and small will increasingly find that algorithms are a crucial part of serving customers effectively. From helping customers find the right products, to reducing fraud, to delivering services more efficiently and cheaply, there are arguably few businesses that can't benefit from a Netflix-like approach.

The challenge of course, is increasingly not data. Thanks to the big data trend, more and more companies are collecting that. The challenge is performing the type of analysis that Netflix-like algorithms perform. With this in mind, the really important trend to watch may not be big data, but rather big analysis.

Patricio Robles

Published 9 April, 2012 by Patricio Robles

Patricio Robles is a tech reporter at Econsultancy. Follow him on Twitter.

2647 more posts from this author

You might be interested in

Comments (3)

Martin Ryan

Martin Ryan, Technology Consulting Manager at Javelin GroupEnterprise

While I agree with the sentiment, most retailers can not devote the resources that are available to Netflix, nor expect to achieve the results that they do.

Most are today content with automated product recommendations from the likes of Baynote and richrelevance. These products (and many others) are rapidly maturing and becoming a staple of basic website personalisation. Retailers that do not yet use them should investigate them as a priority since the return on investment often looks compelling.

The interesting developments are in using these algorithmic personalisation tools with other tools. For example using them with search engine results sets to provide personalised rankings, or with customer service knowledgebases to help reduce contacts per order.

over 6 years ago

Nick Tsinonis

Nick Tsinonis, CEO at RecSys Ltd

The great thing about data is that it does not need to ask questions, just observe behaviours and learn from them and apply these learnings in real-time.

Every business in the next few years will be using algorithms to not only predict users tastes but a mix of personalization, discovery and prediction algorithms allows for:

- higher sales and retention due to personalized marketing via learning through crowd and trend behaviour

- Personalized serindipity assistance. More aha! moments through discovery of new products that customers would never logically look at

- better understanding of customer journey to target customers in different ways in a real-time fashion

- Better prediction of users exhibiting scam, spam, fraudulent and troll behaviour

- New discoveries from user-to-product behaviour leading to new branches of products, that lead product development

... and more

In short, testable and observable behaviours can lead to almost any predictive model being applied.

Nick Tsinonis

over 6 years ago


Cameron Church

For a more indepth look at what algorithms are in place have a good read of this article from How Stuff Works


For the theory heads out there the science (and math) behind the algorithms a good read is on a paper published here


The future is content discovery.

over 6 years ago

Save or Cancel

Enjoying this article?

Get more just like this, delivered to your inbox.

Keep up to date with the latest analysis, inspiration and learning from the Econsultancy blog with our free Digital Pulse newsletter. You will receive a hand-picked digest of the latest and greatest articles, as well as snippets of new market data, best practice guides and trends research.