(containing orderID and productID of each item in the order) it would be
simple. Query the orderDetails table for any order containing the
product ID of the item that the customer is looking at. At the same
time, pull the other items in those orders that are in the same
category. You'd might also want to pull a count of the number of times
that product was bought at the same time so you can rank them from most
to least often bought.
--Jeff
Qasim Rasheed wrote:
> I guess they have some sort of Recommendation and Referral mechanism in
> place. It could be an attributes-based, item-to-item correlation and
> person-to-person correlation (Collaborative Filtering).
>
>
> Thanks
>
>
> Qasim
>
> -----Original Message-----
> From: Tim Blair [mailto:[EMAIL PROTECTED]
> Sent: Wednesday, May 19, 2004 5:22 AM
> To: CF-Talk
> Subject: RE: How does Amazon do that?
>
> > When you look around Amazon.com, and show an interest in
> > something, there's always a part of the page that says
> > "people who bought this also bought that
> > .."
>
> The solution I came up with uses two tables - the product table (guess
> what that is?) and a product_link table. The product_link allows me to
> store, well, links between products in an arbitrary way.
>
> The link table has a 3-way PK of product ID, associated product ID and
> product link "type" - the type is just a numeric field which represents
> the type of link between the products which in my case includes "similar
> too", "related too", "cross sell with" and "also bought". The first
> three are set via the admin.
>
> The final field in the link table is a generic "data" field, which in
> the case of the "also bought" type is a count. In this way I can grab
> the most popular "also bought" products with a nice simple query by
> ordering by the data field.
>
> The "also bought" data is not generated on-the-fly - it could take
> waaaaaaaay too long to increment the appropriate counters for each
> product link when someone checks out, so instead it's done nightly (or
> whenever) on a schedule.
>
> That's just one solution though...
>
> Tim.
>
> --
> -------------------------------------------------------
> <CF_CodingContest mode="judging" newentries="false">
> Maze Solver - http://tech.badpen.com/cfcontest/
> -------------------------------------------------------
> RAWNET LTD - Internet, New Media and ebusiness Gurus.
> WE'VE MOVED - for our new address, please visit our
> website at http://www.rawnet.com/ or call us any time
> on 0800 294 24 24.
> -------------------------------------------------------
> This message may contain information which is legally
> privileged and/or confidential. If you are not the
> intended recipient, you are hereby notified that any
> unauthorised disclosure, copying, distribution or use
> of this information is strictly prohibited. Such
> notification notwithstanding, any comments, opinions,
> information or conclusions expressed in this message
> are those of the originator, not of rawnet limited,
> unless otherwise explicitly and independently indicated
> by an authorised representative of rawnet limited.
> -------------------------------------------------------
> _____
>
[Todays Threads] [This Message] [Subscription] [Fast Unsubscribe] [User Settings]
