Netflix movie recommendations suck!

I’m a huge fan of social networking in general and given my interest in movies, I’ve always been vociferous in my thoughts about movie recommendation sites; esp. if they’re in the social networking realm. I believe Netflix has such a great potential to be a phenomenal social movie recommendation engine / niche social network. However, their current movie recommendation (Cinematch) is no match for a human/social recommendation and maybe that’s why Netflix is on a quest to better the recommendations by 10%. Whoever does that goes home with a million bucks:

Netflix is all about connecting people to the movies they love. To help customers find those movies, we’ve developed our world-class movie recommendation system: CinematchSM. Its job is to predict whether someone will enjoy a movie based on how much they liked or disliked other movies. We use those predictions to make personal movie recommendations based on each customer’s unique tastes. And while Cinematch is doing pretty well, it can always be made better.

Cinematch can definitely be made better. Here’s why?

Picture above: A snapshot of my movie recommendations from Netflix

As an example, recommending a comedy movie, “Ball of Fire” based on my rating for “Yojimbo”, Bicycle Thief, and Rashomon, makes no sense whatsoever since Bicycle Thief was an inspirational tragedy, Yojimbo and Rashomon – two classics from the master filmmaker Kurosawa. Why, Ball of Fire? I may never know but I know there’s a much better way to recommend movies.

All this depends on the database of my movie reviews & ratings. And, guess what — all my recommendations and ratings reside on Flixster which I access via Facebook! I already have 105 friends on that network (who’re also on Facebook). Having said that, I notice that there are not many movie recommendations on Flixster that I value while users of Netflix are likely to be movie fans such as myself. So, here’s what I’d love to see Netflix do:

1. Create an easy way for me to rate & review movies that’ll feed into the Netflix database (already exists on Netflix; not outside of it)

2. Try to collect all this data through a slew of social networking sites. Allow me to tap into my pure-play social network of choice (Facebook, MySpace, Bebo, etc…) — and all those reviews I write on the social network flows back into Netflix’s huge database of viewer tastes. That way, I’ll not only be able to tap into the social network of my choice but also identify Netflix on it.

3. Once I’ve the database of ratings/recommendations in place along with ways to consolidate it, next comes the algorithm to recommend movies automatically. Our taste in movies is a unique combination of factors and I think I’ll be able to describe that better in a future post of mine.

I’ve a few ideas that I’ll be sharing with my friend Michael Rubin (who’s done a tremendous job with Netflix’s community involvement – he even pens their community blog!

What are your thoughts on Netflix’s recommendation engine? Do you think you know a better way to improve their recommendations.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

Comments (



  1. Mario Sundar

    Well, I didn’t know you could tag your films in Netflix, but could be.

    I’ll try to find out from Michael how it may have happened.

  2. DeepTrance

    Interesting! A good part of my queue is based on these recommendations. But then, I haven’t updated my queue in a while. Maybe the changes (and the skewed recommendations) are recent and are based on user inputs and tags. Maybe…

Create a website or blog at