Personalization is definitely everyone’s business

Clearly the giants Netflix and Amazon widely contributed to promote personalized recommendations. The Netflix recommendation is worth a billion dollar per year and Netflix estimates that it drives more than 80% of the video choices. As far as Amazon is concerned, the company, who introduced recommendation pretty much at every step of the purchasing process, claims that its recommendation engine led to a 35% increases of sales.

Should I rely on personalized content recommendation too ?

Definitely yes, personalized recommendation should be everyone’s business, whether your are an e-merchant, a content provider or operating an ad site. More specifically, as soon as you are not able to display your entire product or content collection on the screen of a standard smartphone, you need to implement content recommendation engine.  And a personalized one, so that you can provide your customers with the content that will match their interests. Of course that is if you want to avoid them to experience a long and painful navigation process on your web site.  At a time where the patience of Internet users shrink by the minute, where the volume of content and products increases exponentially, it is crucial to understand your customer and help to find out in a few click the item they are interested in; this is crucial to boost sales and web traffic. Needless to say, the huge volume of data required to be processed, rules out any non-automated solution.

46% of visitors who do not find what they are looking for at the first try leave the site “forrester reseaurch, June 2015” 

Collaborative filtering to the rescue

An efficient way to  implement personalized recommendation is to rely on collaborative filtering. Collaborative filtering relies on the assumption that if Bob and Alice have liked similar items in the past,  they are very likely to like the same thing in the future.  Predictions achieved by collaborative filtering for a given user, do not solely rely on the user’s past history but also leverages the behavior of all users.  This allows to benefit from the “Crowd wisdom”  and capture hidden information about users (i.e. the information which is not available in a database but that is  predicted in real time as customers navigate on a web site. 

86% of e-buyers using personalization technology admit that it influenced their purchasing decision. “Infosys, 2014”

How does that really work though ?

Each user is associated with a profile, that contains the list of items she has bought/liked/viewed… This reflects her interests No need to explore the content nor to analyze the online catalogs. Leveraging the user profiles is enough to identify the most similar users. Their profiles is then leveraged to compute for each users personalized recommendations that can be used to personalize a website, a newsletter or an email.

Thanks to such personalized recommendations, users will be likely to like the items displays for them and to click on them. They might also be delighted to discover items that they are clearly interested in but would have had hard time to extract from a painful web navigation.  This will drastically change the customer experience online as well as the relationship between brands and customers.

80% of consumers say that brands do not understand them as an individual. “Marketing-pro, 2015” 

How do I set up a recommendation service ?

Actually this is very simple. Contact Mediego, which recommendation engine ( coming from the Inria research labs) relies on a powerful accurate and real-time collaborative filtering algorithm. All you have to do is to integrate specific tags on your web site of your emails. Then you only need to select the items that you want to see appear in the personalized recommendations, where you want the personalized content to be displayed and that’s it. From now on, your customers will experience a brand new fully personalized Internet life.