Amazon is smarter than Google
An Interesting article from Yeald.com ...
The principle to sales and support success on the Internet is easy; the more relevant a company’s offer is to the receiver, the higher is the acceptance of the offer.
It’s that easy, in principle.
In practice, things look a lot different. Only a handful of companies have managed to create a successful personalization of their engagement with customers. Amazon personal recommendations and individual composition of the ecommerce site remain the classic example.
Not many other personalization efforts are touted.
Take a look at Google web search personalization effort, and you‘ll likely be disappointed. Event though Google is full of mathematical brains and technical resources, the current status of its personalized search function is way below my expectations.
How come it’s so difficult to build a decent personalized user treatment?
Why does Amazon succeed where Google fails so far?
One of the main reasons is the fact that Amazon understands a user better than Google does. Both companies track user behavior and would be able to track even more behavior than they do today. The user behavior is tracked as very basic, raw data.
What do these basic user behavior data mean?
This is where Amazon is smarter than Google. Amazon has managed to create a layer of meaning to lay over the basic user behavior data. Amazon knows that a purchase relates to a book, a DVD or a CD. Amazon knows that certain CDs have the same sort of music style. Amazon can combine user behavior across product categories and see if there are correlations.
Fundamentally, Google should be able to build a lot more understanding of a user.
In practice, this would require that Google would create a layer of meaning over all user behavior data. But since Google is dealing with a huge and extremely diverse knowledge universe, how could it build a framework that can structure the meaning of 8 billion web pages.
A possible solution is the semantic web. However, so far there are hardly any results known that can live up to a promise of even a basic understanding of the information universe.
The potential of Google’s knowledge is enormous. The company is an advertising machine. And if Google can manage to understand its users, it would certainly find a way to turn this understanding into advertising money.
But this will only work once Google understands the Internet's information universe.
Until then, Amazon remains smarter than Google.
posted by Sujayath @ 10:24 AM
4 Comments:
Amazon has pretty structured and limited data to deal with.
Given data like that a user purchased Pink Floyd, Led Zep, Megadeth, Iron Maiden it is not too difficult to predict that the user prefers music in the Rock Genre. Based on previous user purchase history, it can then predict what other artists the user will like.
But on the other hand, Google deals with larger, more unstructured data. With such data, what kinda personalization can one offer?
I am, therefore, not very sure if the article has its comparison right. It might be like comparing apples and oranges. :(
Cheers,
D.
The basic flaw in the argument you present is that amazon is a store where as google is a searchengine.
Like D. said, you are comparing apples and oranges!!!
If you were to ask for web-surfing patterns of a user from amazon you'd end up with didly-squat but if you were to analyse surfing patterns from google you would be able to tell music/books/gadgets that the user likes. No?
This is in response to Anonymous' arguments -- No it is not entirely an Apple to Orange comparison. Actually, understanding the context in which a user is searching for info is very important. for instance say i am searching for "social networks". i am doing so from the perspective or organization design or a sociology perspective or some such thing. the search engine should not throw up searches like computer networks simply because one one word matches.
this is a hypothetical and may be not a completely illustrative example. but, i hope it conveys the meaning. in fact there have been attempts to incorporate NLP logic into search engines. I remember one Valley startup called Purple Yogi which tried doing that to some extent.
--- ISB Alum (2003)
This comment has been removed by a blog administrator.
Post a Comment
<< Home