 |
|
|
 |
 |
March 2004 |
  |
The Missing Metric
Whether a web site is successful depends on the definition of success
By JJ Hollowell and Jeff Lattner
Among the advice that consultants are freely dispensing
to retail clients these days: Deliver shoppers from search results directly
to a product page. Don’t make your customers who are looking for a
particular product do another search for it when they arrive at your site. Extremely good advice, but sometimes more difficult to
put in place than a retail operations or marketing chief might suspect. The
problem comes when retailers take steps to create a personalized shopping
experience.
Most e-commerce managers today would agree that a
retail site’s objective should be to be user-friendly, easily managed
and profitable. Often, to achieve their goals, they implement session IDs
or cookies to personalize the visitor experience and they purchase an
enterprisewide content management system to deliver a single experience to
users. But managers at a fairly large e-commerce site, with a
few hundred products, recently learned the hard way that focusing on
personalization was hindering their ability to attract customers. Upon
close examination of their operation, they discovered that session IDs,
cookies and the content management system that so perfectly served visitors
once they got to their web site were actually obstacles when it came to the
web site being indexed by the leading search engines. They had learned the
hard way that those investments often can have a counter-intuitive effect
on a web site’s profitability. Tripped by scalability The problem stems from the explosive growth of the web
and the consequent demand for qualified search results that have forced
search engine technology to scale dramatically. Search engine spiders crawl the web gathering HTML and
following links, retrieving and storing information in a database to be
analyzed later. But session IDs, cookies and content management systems are
challenges for search engine spiders. Spiders have a hard time indexing
dynamic URLs that contain session IDs for fear of indexing duplicate
content.
Here’s how the problems develop: Retailers want
to identify customers as they move through a site or when they return to a
site. Thus, their content management systems, or in some cases, their web
servers, generate session IDs that are transferred to an end user via a
cookie. Once the cookie is deposited with the user, the retailer’s
system will recognize this specific user when he returns to the site. This
is beneficial for personalizing the end user experience and for the
retailer to learn about customers’ shopping habits and patterns. But
if the end user’s computer will not accept cookies, the session is
tracked by a session ID that is encoded within the URL string. However,
return sessions cannot be recorded when using session IDs because the
individual receives a new session ID each time he returns to the site.
That creates a big problem for search engine spiders
because they do not accept cookies. And since spiders crawl an individual
site several times, they get a new session ID within the URL string each
time. This gives the search engine spider the impression that it is
crawling new pages, when in reality it is just indexing the same pages over
and over again, creating duplicate content. Big trouble
To a retailer, that could spell big trouble. If a site
is not completely crawled, the amount of content available is reduced for a
search engine to determine the relevance or importance of a site to a
particular search query. This, in turn, dramatically reduces the number of
times a site will be found by spiders and will ultimately reduce the number
of visitors resulting from search.
The good news is that, by understanding the issues
between content management systems and search engine spiders, web site
owners can implement strategies to maximize benefits from both resources. A
search engine optimization agency can provide recommendations on how to
disable session IDs specifically for search engine spiders, enabling a
retailer to realize the benefits derived from the content management
systems. For instance, search engine optimization agencies can help
retailers program their content management systems to determine when a
computer accessing a site is a spider. It’s a fairly easy process
that involves detecting whether the device accessing the site is using a
consumer browser, such as Internet Explorer or Netscape, or if it is using
some sort of spidering agent, which would indicate that a search engine
such as Google is doing the looking. When the system detects a spider, it can both disable
cookies and prevent the session ID from being injected into the URL. Some
content management systems have that functionality built in; all it takes
is a change in programming to activate it. Others don’t, and that
will require re-writing code to include that functionality.
Even though search engine companies are trying to find
ways to accommodate session IDs, solutions don’t seem imminent, so
search engine optimization companies recommend removing the potential
issues from session IDs immediately to avoid any drop in search traffic.
When hiring a search engine optimization agency, retailers should make sure
the agency has the knowledge and experience to remove session IDs from URL
strings, while enabling the retailer to realize the benefits derived from a
content management system. For instance, retailers should ask agencies
about their specific experience in working with the retailers’
content management system, either activating the appropriate procedures
within the system or writing code to create the necessary enabling or
disabling. Know the whys Before you talk to a search engine optimization agency,
make sure you can tell the agency why session IDs are implemented on your
site in the first place. If session IDs are not essential, for instance, if
nobody at the company is making use of the data that session IDs generate,
the best recommendation is to remove them completely. This will eliminate
the session ID issues and will allow spiders access to crawl your site. For
e-commerce sites that require cookies for checkout, first look at
installing technology to detect whether an end user has cookies enabled
(less than half of e-commerce content management systems/servers have this
capability today), then create a pop-up window for users who don’t
accept cookies, prompting them to do so. This is a very straightforward
step that can be taken with a simple JavaScript prompt, and will eliminate
the need to generate session IDs. If session IDs are essential, for instance, if you want
to keep information only for that session not for later use, look for other
ways to gather the analytics they provide. For example, while log file
analysis has taken a lot of bashing lately because it’s hard to get
usable information fast from log file analysis, it might be a good measure
to take until you’ve resolved the search engine issues related to
session IDs. Whatever approach you take, session IDs should be dealt
with immediately to remove obstacles to search traffic. The benefits to be
gained from higher visibility can translate into the profits you need to
make your site even better.
JJ Hollowell is chief technology officer and Jeff Lattner is optimization consultant for search engine marketing company icrossing Inc. They can be reached at jj@icrossing.com or jeffrey.lattner@icrossing.com. 
|
|
 |
|
