Predictive Analytics and Why Companies Are Stalking Their Customers

Tomorrow is all about Big Data and how best can you handle it. See, companies don’t need more data. Most medium to large companies either have the data or ways to get it easily available. The problem is – most of them don’t know how to handle it.

Here comes the boom: Predictive Analytics is *the thing* nowadays. Long gone are the days when merely registering data, processing it and acting upon the findings in the next fiscal year was enough. Right now the fastest growing companies register data, analyze it and respond to it in real time.

Predictive analytics and predictive personalization – how do they work?

Hunch Recommendation
Hunch, now part of ebay, offers personalized recommendations based on personal infromation

We all leave trails behind. Our shopping habits, our marital status, our social groups, the shows we watch and gadgets we buy – all these and much more are trails and they are in some database, somewhere. Using this data, or whatever is available at any given moment, predictive analytics software can determine our future actions through two types of programmed responses (it’s a little bit more complicated than that, but you’ll get the picture):

1. Rules Based Personalization – “If this than that”. Basic personalization. Ex.: Customers click on an ad, enter our website and we can determine they are from New York. Let’s show them our stores in New York first. They click on our product catalog, select the high-priced products. Bang! We now know they have a medium to high income. This kind of responsive personalization does not really make use of any kind of predictive analytics. It just reacts to actions. It does not try to predict them. This is a job for…

2. Predictive Personalization – this is something we, humans, can do easily. Machines, not so much. Let’s say our sports store has a sales person with a decent IQ who’s at least a little bit interested in the customers checking out the merchandise. He notices customer X has tried on at least a dozen of sports shoes in the last hour. He walks to the customer and asks him “Hey, can I interest you in this brand new snowmobile? It’s 10% off. Oh, wait that is stupid. That just what old-time ads would do. He would actually ask the customer if he can help him find some shoes that fit and look good. That’s basically what Predictive Personalization is all about: 1. Analyzing the data real time / 2. Using context to pinpoint the best potential recommendation and 3. Personalize the output.

In case you were wondering – yes, there’s a little bit more science to it but the previous example shows what the buzzword stands for. If you are interested in the subject or you’re a future Predictive Analytics Expert you can have a look at “Personalized Recommendation on Dynamic Content Using Predictive Bilinear Models”, on how Wei Chu and Seung-Taek Park of Yahoo Labs used Predictive analytics to recommend better content on Yahoo’s front page.

Why companies use Predictive Analytics to stalk their customers?

You know why Facebook stalking is so easy? Because people want other people to know about their interests. The Millennials, the digital natives, generation Y – they are today’s youth and they are born and living online. They offer their info, they share their interests, they make their photos public. No more mass message. Each and everyone expects to be treated as an individual.

Companies that do not “stalk” their customers are going to be left behind: Amazon is personal, Facebook is personal, Google is personal. Most of the top online retailers are personal and they make customers’ shopping experience unique.

How about offline? Yes, 5 years ago we couldn’t have had any kind of Predictive Analytics or Personalization offline but the iPhone changed that. Now smartphones fill the gap between the data stored online and offline activities. Companies are now tracking consumer behavior through mobile activity and make use of predictive analytics to address individuals needs and wants … well .. individually.

Acting on data is not enough anymore. It’s acting on data NOW that’s important.