NONE: Re: ONLINE-ADS>> The Amazon.com model
Re: ONLINE-ADS>> The Amazon.com model
Les Morgan (lmorgan_at_IX.NETCOM.COM)
Thu, 24 Jul 1997 10:58:20 -0700
Jason:
This is in response to the posting Amazon.com. I'm just in the process of
bringing a new online bookstore up at one of the sites I manage and I've
been looking at various bookstore design models. The new bookstore we're
doing has a narrow specialization for death and dying, with the main
audience segments being health care professionals in hospice settings, and
health care consumers who are impacted by death. Some of the items we plan
to carry are hard to find, so Amazon.com has been a good choice due to the
branding and inventory.
I'm not sure if the following statement is correct:
>treknews_at_juno.com (Jason T Ellis) wrote:
>I've tracked it down to being a problem with Amazon.com's system ... once
>a user clicks through to a book, they're then on Amazon.com's site. If
>they go anywhere else... I loose the sale by default. They're only
>tracked to that one single page.
According to the Amazon.com sales activity reports, sales tracking is not
limited to the single page they first visit. My understanding of
Amazon.com's cookie-based visitor tracking is that all sales made during a
single visit to Amazon.com go to me, even if the visitor moves to another
page at Amazon.com. If they leave Amazon.com, go to another site, and then
return to Amazon.com, the sale to me is lost unless they used their PAGE
BACK button to return to Amazon.com, which would not update the cookie last
deposited on their system by Amazon.com as a result of the original visit
from my site.
This opinion is based on my reading of the following definitions, which are
taken from the Amazon.com quarterly statistical report for my site:
>A "Visitor" is a person who clicks on book links from your site,
>and is counted as 1 visitor (above) regardless of the number of
>different titles they click on. We keep track of this by watching
>their shopping basket ID, which remains the same for every book they
>click on.
>A "Hit" is any person clicking on a book link, and each click is
>counted as 1 hit. If the same visitor clicks on 5 different titles,
>we record 1 visitor and 5 hits. Therefore, you should expect the
>number of visitors to be lower than the total number of hits.
In other words, "visitor" ids persist through a shopping session.
I'd like more detail on what you know about how this actually works and
what the problem is that you have noticed -- help me to understand what
you've discovered.
There's a separate design issue here, which is how to design your online
bookstore so visitors do their browsing at your site and only go elsewhere
(if at all) to execute the purchase. It would also be nice to be able to
offer books from more than one source -- you can offer items from
Amazon.com unless they are only available elsewhere, if the discount is
better elsewhere, etc.
To position the bookstore to do this, we're experimenting with the tactic
of creating a separate book review page for each book. These pages are
interlinked with other content areas of our web site, and with the main
bookstore front end, with a goal of trying to get the user to make a buying
decision on our site, not at Amazon.com, and only go to Amazon.com once
they've seen enough to be sure they want to make the purchase. Because
each book review is treated as a separate sales item, we could have
relationships with multiple booksellers.
We don't have a sitewide single shopping basket, but I'm not convinced that
this is needed because sales reports show that most people only order one
book per visit (this assumes that the tracking method keeps track of
visitors by session, as stated above). If the Amazon.com tracking report is
inaccurate, as you suggest, they my site design is based on a false
premise! :(
To increase book targeting, our book review pages are also integrated with
our general search engine, which includes web sites and other things, so
book suggestions are presented to the user in response to keyword and
category searches. Most book pages are called up as a result of these
searches, and we have relatively few visitors to the actual bookstore hub.
We offer the search engine for remote installation on hospice sites as an
interactive feature for them.
I'm very interested in feedback on this design approach, treating each book
as a standalone item and integrated with a general search engine. We're
still in the early stages with this bookstore, and want feedback on the
beta version before ramping up the inventory.
Main site: http://www.growthhouse.org
Bookstore hub: http://www.growthhouse.org/books/books.htm
Search engine: http://www.growthhouse.org/search.htm (try a keyword search
on the word "grief" or on the word "hospice")
-----------------------------------------------------------
Les Morgan lmorgan_at_ix.netcom.com
The Vision Team http://www.visionteam.com
415-863-1617 voice
415-863-2114 fax
----------------------------------------------------------------------
This week's Online Advertising Discussion List sponsor:
The New York Times on the Web targets ads with EXCEPTIONAL PRECISION
due to user-supplied demographic information on 1.7 million registered
users. Find your audience. ---------->>> http://www.nytimes.com/adinfo
-----------------------------------------------------------------------
Online Advertising Discussion List To Unsubscribe send UNSUBSCRIBE
http://www.o-a.com/ to online-ads-request_at_o-a.com
