Skip to main content

Recommender engine part 1: What it is and why you need it


Websites of all types should consider these before discounting the use of a recommender engine.


The future of the web is helping people shop and experience sites in a uniquely personal way for them; whether the website is e-commerce, media, social network, or search engine, this is a win-win situation for both the users and the site owners (or store owners, in the case of e-commerce).

One way to do this? Utilising a recommender engine. A recommendation engine, also known as a recommender system, is software that analyses available data (user behaviours, product properties, etc) to make suggestions for something that a website user might be interested in, such as a book, a video or a job, among other possibilities.

Amazon was one of the first sites to use a recommendation system.


Recommended product from Amazon website.

Websites have a lot to gain in using a recommender engine, including:

Maybe change Search is hard for users, discovery/exploration is easy to Search limits the users to what they already know while discovery/exploration offers more and is easier and more convenient for users to make choices.

Increased sales
Using personalised recommendations are proven to increase sales. Amazon, for example, reports that 20-30 per cent of sales come from personalised recommendations. Same as with these other examples:
  • Netflix: 2/3 of the movies watched are recommended.
  • Google News: recommendations generate 38% more clickthrough.
  • Amazon: 35% sales from recommendations.
  • Choicestream: 28% of the people would buy more music if they found what they liked.



Better user experience. 
Show users personalised and more relevant products while their on the site and off (via email). Also give users tools/games to discover more of what they like. Search is hard for users, discovery/exploration is easy.

Increased traffic/page views. 
By giving users more discovery tools you will receive more traffic and higher product discovery potentials.

Improve customer retention. 
Also means higher lifetime customer value. The more users like your site, the more the buy, the higher they are likely to return. Also use targetted and personalised email campaigns to bring them back and keep making them happy.
Recommendation engines are increasingly becoming popular and for good reasons, too. Have you used a recommendation engine? What other benefits have you experienced?


Comments

Popular posts from this blog

PHP json_encode return empty array [] instead of empty object {}

Problem: Get below array for example.  $status = array ( "message" => "error" , "club_id" => $_club_id , "status" => "1" , "membership_info" => array (), ); This array will be encoded in json format echo json_encode($status); This function return json: {"message":"error","club_id":275,"status":"1","membership_info": [] } Notice the empty array [], it is not consistent, it should be an empty object {} {"message":"error","club_id":275,"status":"1","membership_info": {} } The cause: This problem is caused by the called function array(), which yields an empty array [] Solution: There are 2 solutions for this issue: Solution 1: Call new stdClass instead of array(), stdClass generates an empty object {} Solution 2: The above solution is complex in case there are plenty of arr...

Command Prompt VS 2010: Fix MSVCRT.lib(MSVCR100.dll) : error LNK2005: xxx already defined in LIBCMT.lib(xxx.obj)

One of my task is to use Command Prompt of Visual Studio 2010 to evaluate the assignments of students. This assignment is about pointer and memory management in C++. The students are required to use new and delete operator in C++ to manage the memory and they need to ensure that the number of new callings must be equal to the number of delete callings. I overload the new and delete operator, build it into a static library (.lib file) and require student to use this .lib file when implementing the assignment. There is no problem when building the .lib file in VS 2010 project, but there are errors when compile the source code if I link the source code to the above static .lib file. Here is the format of error: MSVCRT.lib(MSVCR100.dll) : error LNK2005: xxx already defined in LIBCMT.lib(xxx.obj) Fortunately, I found 2 ways to fix this error on the internet. The first one is to remove the default lib when compiling the source code, it's libcmt.lib, because this libcmt.lib alread...

Fix cross domain in ajax request

Follow this sample ajax request $.ajax({    "type": 'POST',    "url": url,     "crossDomain": true,    "data": params,    "dataType": 'json',    "async": false,    "success": function(data){ done_fn(data, is_test);    },    "error": function (jqXHR) { var data = $.parseJSON(jqXHR.responseText); $("#results").text(data.message);    } }); The keys are crossDomain: true and async: false Hope to help many people.