Big data rules but, Netflix’s Secret Special Algorithm is a Human

hawaii kane

Imagine the scene.
The great and the good of the Polynesians gather in front of the cave where dwells Kane the great god of wealth and power eager for a sign of some sort that will guide them in their quest.
In the cave is the devious son of the Islands only coconut farmer and with a deep, practiced voice he says, “Eat coconuts every day and you will be rich beyond your wildest dreams”.  Nobody in the audience hears or sees anything, but what they desperately want to hear and see, nor do they want the trouble of a bit of thinking and deciding and  soon they all drop to their knees in worship and the party goes on until dawn.

Guess who got rich and powerful beyond their wildest dreams?

In this blog,  I talked about the powerer of personalisation and went to some depth about the possibilities and limitations and specifically we discussed one of the acclaimed  leaders in personalisation, Netflix. Most people are familiar with them and believe they do a fair job of making recommendations.

Imagine my chuckles and many peoples amazement to read that;  “Netflix’s Secret Special Algorithm Is a Human

Yes, Netflix has used their customer insight to fund a movie and the movie  was watched by lots of people (duh!) and they made this clever decision that was watched by so many people because of the remarkable data insight they have.

No, not quite!   Here’s what happened.
Netflix asked an individual who has probably demonstrated this ability to them in the past to decide what their customers would indeed watch.  Having funded the movie, they then threw the might of all their considerable marketing ability and access to their own customers into making sure they did indeed watch and talk about the movie.

 polynesian big data  conference

DataCon 2015

It’s one thing knowing a few things about how to make pieces of data comparable and physically compatible and even knowing how to get it out of all the competing pieces of technology and into one place and occasionally someone may even be able to preserve scraps of the context in sort of usable form, but it’s quite another to assume that this explicit data is knowledge let alone intelligence.

Thee are many clever people working on development of tis higher level of machine intelligence and with a very long way to go. In the meantime it does no harm to classify knowledge in tis kind of fashion:

Explicit – mostly just data and not fit for purpose

Implicit – Enriched by comparison to other knowledge that supports developing good hypothesis.

Tacit –  Enriched by the experiential knowledge of people and experience and ready for use with caution and consultation. Clearly this is what Netflix have done  quite successfully and hopefully others will learn from.

Coconuts are rich and pleasant food and there’s plenty to be said for eating them, but there’s a limit. Clever collection use and interpretation of data is a powerful thing and nobody should be ignoring it, but I have to laugh out loud at some of the stuff that is written, often by people who have recently graduated form a typewriter.

For further reading see:

Using personalisation cleverly to grow your customer relationship and keep your sanity

Are you wondering if a recommendation engine is the next big purchase for you?

Give big data the heave-ho for now and get a birds-eye view of your business without a single nosql database.  

About the author

2 thoughts on “Big data rules but, Netflix’s Secret Special Algorithm is a Human

  1. Since Google is already a dependable and established brand as a
    result of other services such as Gmail, Google – Maps and Google – Places, people are
    more likely to trust how the Google+ platform is a helpful
    tool as well. Mega sites, however, are fantastic for Ad –
    Sense. The concern is that without people jumping up
    to speed, sharing and checking into Google Plus regularly;
    that only leaves brands to market with brands.

Leave a Comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.