Google's Meaning Layer

READERS How Google builds systems to enable the extraction of meaning from distributed information.
Betrifft: AMAZON COM Customer Relationship Management Direct Response Marketing GOOGLE Internet Search & Directory Media Retailing & Rental Strategy

by OWT
The author is registered user with YEALD. To get your own profile please become a Recognized Writer. more...


Yes, Amazon does have a very good system to extract meaning from seemingly disconnected information, but that is because the system is layed out to allow that to happen.

With Google, things are layed out differently. Advances in search technology allow for more meaning to be picked out of the stream of information. The bigger the system the better the prediction. So where does Google get their meaning? From an analysis of connections for one. For this we need to take a closer look at what Google already knows and might know in the future.

First they know what I search for, especially if you use the personalised search they are offering at google.com/ig. That is a lot of meaning right there if you aggregate it. Above that they know which links I click and if that site as Google AdSense on it, they know that I ended up there, how long I spent, and if I went to further sites. If those further sites also have Google AdSense they can track my entire path. Due to the login I have at google.com, any cookie they set cannot really be deleted consistently because it will always be reset with the right ID when I log back in.

If I click on an AdWords link, either on Google or on a Network site, then they know what that click was worth, and as some AdWords users use their tracking system to integrate further steps along a sign-up or sales process into the AdWords system, they potentially know that I bought something and might even know what it was worth for the advertiser.

If I searched for Sony, TV, HTDV and other items and then clicks a HDTV ad, that was linked to HDTV, Buy, Big, Crisp, ... they know I am looking for a HDTV TV with a big screen and if I buy it they know I have it. They also know through which shop as that name is attached to the AdWords account.

When Google Base launches, they will add a whole layer of meta-information to the mix as they seem to have set up the service in a way that you need to add meta information like this is a house I sell for 350.000 EURs in Cologne, Germany, with 350 square meters of ground.

Once they grow their own WiFi network which is starting in San Francisco, they know exactly what I surf to. The same is correct for the Google Accelerator, which has just relaunched. Through Google Mail they know who I talk to and possibly what they search for and do. Aggregating this allows them to find out about my community of practice and general interest.

Yes, it is not as clear and easy as for Amazon, but through the diversity the amount of information is a lot more rich, if analyzed correctly. And that is why they need the large amount of servers they have and have developed their own API and programming language internally that is there for one purpose only, analysis of huge amounts of data distributed among thousands of servers.