The Real Secret to Amazon, Netflix, and Spotify's Success

To clear any misconceptions, this stream isn’t about some covert Tech perk like nap-pods or on-campus bikes that exponentially increase productivity amongst employees.

I’m going to be talking about an important product strategy that Amazon, Netflix, and Spotify, not only have in common - but are at the core of what makes them so unique and successful.

Let’s begin with explaining The Netflix Problem.

How do you recommend the perfect set of movies to a user to make sure he/she loves what’s offered next?

Here’s what you think you watched:

  • The Notebook.

What Netflix knows you watched:

  • A movie with Ryan Gosling and Emma Roberts

  • A romantic comedy

  • Something directed by Nick Cassavetes

  • A movie inspired by Nicholas Sparks

  • A 2 hour 4 minute movie

  • A movie with the tags: War, love, romance, music, South Carolina, 1940s

Not only all that, but Netflix knows all of this and so much more:

  • When you paused the movie to take a break

  • What scenes you rewatched

  • What parts got boring and you skipped over

  • How old you are

  • Where you stay (not in a creepy way?)

Knowing all of this, Netflix tries to give each and every movie on it’s platform a hidden predicted rating for each and every user.

On a very simplified level - Netflix knows the IMDb rating of The Notebook, what type of users watch the movie on their platform, whether I watched it without interruptions, and whether I liked it.

As a result, Netflix searches for movies that fit similar checkboxes and gives When Harry Met Sally an 8.5/10 (for example) as my personal hidden predictive rating.

And when I’m done with my first movie, I’m dutifully prompted to check out another movie that I’m likely to enjoy.

Not an endorsement of Kevin Spacey

Not an endorsement of Kevin Spacey

What Netflix is doing, in other words, is guessing what movies I’ll enjoy, before I even watch them.

I know what you’re thinking right about now.

It was Emma Stone, not Emma Roberts in the movie, as I mistakenly mentioned above.

And you’re right - I’m just keeping you on your toes.

But, let’s take a look at how the same idea can be found at Spotify.

They have managed to become one of the leading streaming services in the world, with over 170 Million users (70 Million of whom are premium). They’ve constantly rivaled streaming services offered by the likes of Apple, Tidal (Kanye and Jay-Z), Soundcloud and come out on top.

They solved an even harder version of The Netflix Problem.

Spotify uses even more granular data - when you skipped a song, what type of songs you added to a certain playlist, how many times you listened to a song on repeat, etc.

The primary way Spotify recommends new music is through finding people similar to you - your Facebook friends, people in the same area as you, people with similar songs in their playlists.

A visual on what this looks like.

A visual on what this looks like.

Spotify wants to predict what songs you’ll like before you even listen to them.

I’ll explain why this all so important in 60 seconds - but first, let’s take a look at Amazon.

Amazon showed a 29% increase in sales Q/Q, primarily due to the implementation of its own recommendation engine. When you click on an item, Amazon is quick to recommend items that are “frequently bought together”. As you scroll to the bottom of your selected item, you can find a “featured recommendations” carousel of products. “Customers you bought Product X also bought” is just another example.

recommenations amzn.jpg

Amazon wants to guess what products I want to purchase before I even know I want them.

Okay - these 3 services that I use constantly are always predicting things about me - why is this important?

Well - recommendations are what allowed Spotify to beat out Apple Music. Recommendations keep users like you and me on the application, which is an important metric for companies.

With the proper recommendations - you listen to music on Spotify for hours, and you decide that you really enjoy it. You keep getting more and more songs that you enjoy, and you keep contributing to ad revenue for Spotify. Alternatively, you decide to just shell out the few dollars a month and you become a premium revenue source for Spotify.

A similar concept applies to Netflix and Amazon - you enjoy the movie/product you watch/receive, and every subsequent recommendation simply serves to solidify that you’re paying for a worthwhile service.

Recommendations are fueling the revenue streams for these companies, and many others like them. Knowing what you’ll like before you even do, is a powerful tool.

What accompanies this ease-of-use, and largely personalized service, is a collection of user data. All these companies build profiles around you to best determine who you are, what you like, and what you will like. There are arguments on both sides - if you don’t utilize user data, your business is susceptible to becoming irrelevant. If you do use it, you have to be careful what data you store, and how you store it.

But either way, recommendations are a critical reason behind why these companies are successful.

Next time you’re recommended a song, or product - think about the real reason you’re getting it, and the amount of thought that went into trying to perfect the recommendation.