Alexa Web Search Platform
Wow. I was checking out the Amazon web services homepage for some affiliate stuff and noticed there were some new Alexa related web services. Turns out Alexa is opening up their crawler and index to anyone who wants to use it, for a cost. The new service is called Alexa Web Search Platform.
So how much does it cost?
$1 per cpu hour ($0.50 for reserved but unused hours)
For every hour that a computer is dedicated to your use, your account will be charged $1. Reserved but unused time will be billed at $0.50 per hour unless 1) the reservation is canceled at least 48 hours prior to the start of the reservation or 2) the reservation has not started and is canceled less than one hour after it was made.
$1 per GB/year of user storage
A multi-terabyte storage system, the ‘User Store’, is available to store user applications, source code, and data processing output. Your total storage is measured twice hourly and billed at a rate of $1/Gigabtye/year.
$1 per 50 GB processed
Pay $1 for each 50 Gigabytes of internal data transferred to or from your reserved computers. This bandwidth may result from data transferred for processing, output written to the ‘User Store’, any communication between your reserved compute nodes, and regular system administration activities.
$1 per GB uploaded/downloaded
Pay just $1 for each Gigabyte of data you upload to the Platform or download from the Platform to your home or office network.
$1 for every 4,000 user-published web service requests
Use our self-service publication system to publish new search services. Pay $1 for every 4,000 requests to those services.
So what exactly can you do with the Alexa Web Search Platform?
Anything you would be able to do if you actually owned Alexa. You can crawl documents, process documents, create your own vertical search engine, create new web services, or store large amounts of data. This opens up the possibility for businesses to be built upon this platform. You can bet Google and Yahoo are going to think about releasing something similar in the future.
How much would you pay to have your own copy of the entire internet? Or be able to analyze whole markets without having to buy the resources you would need to process all that information? This is a big deal and a great way for Amazon to grow a big user base rapidly.
Here’s an example of what can be built on the platform: http://photos.alexa.com/
Right now the only code sample Amazon has is of a Java REST Sample making a request, but I’m sure as the news spreads there will be many more examples.
For more information, take a look at the Alexa Web Search Platform User Guide. Over the next week, I’ll try and write up a more comprehensive post on how to use the web service.
What does this have to do with search engine marketing?
Well for one, you could build a tool analyzing every single page that links to your site – including anchor text, subject, the sites that link to those pages, alexa traffic of those sites, etc, etc – for a very small cost to process. Personalized link network analysis of client sites would be a great service. Right now our current tools are based on what Google and Yahoo return back through their APIs. The information we get back through the APIs may not be exactly the information Google and Yahoo use to calculate rankings. With access to your own full blown search engine, we have access to the same information the big search engines do.
With this service you can not only see who is linking to you, but who is linking to those sites – a real link network. You can filter out by criteria important to you. Let’s say you only need data on search engine marketing sites, you can filter for that. Or maybe you need search engine marketing sites with a certain level of traffic and amount of incoming links, you can filter for that. This is a powerful tool.
I can also see many mashup type applications being built off of this service – individualized search engines, vertical search engines, specialized reports, etc.
Update: Battelle had the inside scoop earlier today.