|
Web Traffic Analysis Software
An Industry
White Paper
Written by: Mike McClure, V.P.
Marketwave
206-682-6801
www.marketwave.com
Executive Summary
Web Traffic Analysis Software is a relatively new
market segment, resulting from the incredible growth of the World
Wide Web over the last decade. As Web sites mature, companies
are beginning to get more sophisticated about how they use this
medium and want to know how to increase their return on investment
(ROI). This paper reviews the trends in Web traffic analysis
and presents marketing research about what is most important to
users in this segment. It also segments the market based on these
user needs. It concludes with recommended review criteria for
evaluating tools within each market segment.
The Need for Web Analysis Tools
No matter how you measure it, the growth of the Web
is phenomenal. The number of Web users is now measured in the
tens of millions while the number of Web sites is now measured
in the millions. In fact, the Web is growing so quickly that
it is useless to quote growth statistics, as no one can consistently
quantify or agree on the actual figures!
Regardless of the actual numbers, it is clear that
a lot of time, attention, and money has been spent on the Web
by companies wanting to get involved in cyberspace. Very few
of them have much of a feel for their payback on this investment.
Much of that has been due to the incredible hype and fast growth
surrounding this technology, combined with the low cost of experimentation.
All one really has to do is develop a Web site and revenues will
shoot through the roof, right? Well, not exactly…
The Web is fundamentally different that most other
marketing venues. In traditional marketing, it's relatively easy
to construct a profile of your target audience in terms of demographics
and psychographics. With the Web, however, you don't know much
at all about the visitors to your web site - as most visitors
are anonymous. They come and go from your site without
much of a trace. The more sophisticated marketers try to collect
demographic information anyway, so that it can be used to target
future content and messages. This is typically done by offering
visitors something of value in exchange for filling out a form.
This need to better understand visitors and what
they do on a Web site is behind the increased demand for Web traffic
analysis software. This software provides behavioral data about
visitors, including the types of companies that visit your site,
where they came from, and what they do when they get there. As
the number of internet users and sites increase, as the number
of page requests increases exponentially. Web traffic analysis
software can help to draw conclusions from this mind-boggling
volume of behavioral data.
The ultimate goal is to combine this behavioral information
with traditional demographic and psychographic information. This
allows measurement of what people say, how they feel, and most
importantly, how they actually respond (which are not always consistent,
or even logical). This information is the foundation of personalized
one-to-one marketing techniques, allowing a business to target
specific audiences with customized products and services that
directly solve their problems. We call this Web-mining (i.e.
using Web behavioral information as an integral part of your one-to-one
marketing).
Typical Initial Phases of Web Site
Development
As companies get more electronic experience, their
Web sites continue to grow in both sophistication and importance
to their organization. Talk to any Webmaster about their site
and you will hear "We're not quite done yet, check back again
next week…". The point is that most sites are going
through a constant evolutionary process. The best description
of this process we have found comes from the Patricia Seybold
Group, which is roughly:
Phase 1: Brochureware
In this phase, an organization's
primary focus is simply to establish a web presence. The driving
force behind an organization's entry into the online world is
simply to "be there". This "first-generation"
web site typically duplicates information about the company and
products that already exists in paper format. Companies tend
to look at their site as a cost effective replacement to the mounds
of paper (brochures, spec. sheets, corporate information, annual
reports, etc.) that they already produce too much of.
The benefits in this stage are tangible, but limited.
Cost savings are realized in terms of reduced production, printing,
and mailing costs. In addition, getting on-line broadens a company
reach into the world. The majority of today's corporate web sites
are still in the brochureware phase. Other forms of Electronic
Commerce were simply not part of the initial specification or
thought process for the site.
Phase 2: Customer Support
This phase attempts to
make the web site a more compelling, bi-directional source of
information for customers. It is also one of the least publicized,
but highest value stages. Internet savvy customers are quite
happy to help themselves to this king of information. This includes
on-line support documents, order status information, interactive
product demonstrations, product registrations, and links to other
valuable internet information. It is very valuable to this community
to obtain information without having to wait for someone to answer
the phone only to find out that they are only two days into the
job and really can't help much. It is also useful when information
is required and you are located in different time zones, or during
holiday periods when people are not always available.
National Semiconductor
(http://www.national.com) and Federal Express (http://www.fedex.com)
are both good examples of sites with extensive customer support
capabilities built in.
Phase 3: On-line Transactions
The next typical stage is to automate the flow of money - a logical
step after providing prospects with all the product and service
information they need on-line. With the number of spot purchase
decisions made in the world, the ideal time to present a product
offer is when the prospect is actually on your web site, interested
in what you have to offer.
While we don't expect the Web to completely replace traditional
hands-on selling methods, significant revenue is already being
recognized through the web. Dell ( http://www.dell.com/ ) and
Travelocity (http://www.travelocity.com/ )
are generating millions of dollars in revenue over the web. Also,
contrary to what is often printed in the press there are thousands
of smaller known companies doing business this way, who are profitable
and thriving. Examples include 1-800-Batteries ( http://www.1800batteries.com/ )
and Greet Street ( http://www.greetst.com/ ).
Phase 4: Personalization
Technology has now made it possible to easily collect and use
information about your visitors and use it to better target marketing
messages. Arguably, the most famous example of this is Amazon.com
(http://www.amazon.com). By building
a customer profile databases, Amazon is able to suggest specific
books that may be of interest to the repeat visitor.
Again, there are many smaller companies using these techniques
to target marketing messages, who don't get nearly the recognition
they deserve. Examples include Freeshop (http://www.freeshop.com/)
and Nth Dimension (http://www.sharpshopper.com/).
Phase 5: Community
This stage is the result of a site that has established trusting
relationships with their visitors. Sites in this stage are heavily
trafficked and often talked about on and off line as providing
real value and content. Often, there are chat rooms and user
groups that are formed among visitors with common interests and
issues. The most famous example of this is demonstrated by Cisco
(http://www.cisco.com).
The Ultimate Goal - eCommerce/eBusiness
While the phrase "electronic commerce" is now in mainstream
use, it is one of those elusive terms whose definition seems to
vary depending upon who you ask. eCommerce used to be synonymous
with using Electronic Data Interchange (EDI), the standard format
to automate monetary transactions. This format is typically used
by large corporations engaged in millions of transactions per
year, but has yet to enter into mainstream use.
Today the term eCommerce has been expanded to include other business
processes that can now be enhanced electronically. Some example
definitions include:
"A set of strategies, processes, tools and supporting
infrastructure to transact business electronically"
- Ed Edwards, The Boeing Company
"Building and sustaining business relationships with customers
electronically."
- Patricia Seybold, The Patricia Seybold Group
"Electronic commerce is the application of technology
toward the automation of business transactions and workflows."
- Ravi Kalakota, Author of "Electronic Commerce, A Manager's
Guide"
IBM has made recent attempts to redefine this concept as eBusiness,
to help avoid the confusion and differentiate themselves among
them many companies that are laying claim to this emerging market.
"eBusiness is about transforming key business processes
by using Internet technologies"
Regardless of the definition, it is clear that a lot of attention
is being paid to this concept, and for good reason - it offers
compelling business payback. eBusiness promises to significantly
improve the "friction" associated with business processes
(or transaction cost, as referred to by economists). Ultimately,
this reduction of friction will enable smoother transactions between
buyers, intermediaries and sellers, which is a significant cost
in the value chain of almost every industry in the world. We
choose to use the term eBusiness throughout the rest of this paper,
to indicate that we are referring to the more general concept
of automating business processes.
With true eBusiness, it is business requirements, rather than
technical elegance, that drives the specification of the web site.
Web sites become increasingly personalized to the constituents
they are trying to support, namely, customers, investors, and
employees.
The most easily measured benefit from eBusiness is the cost savings
associated with allowing customers to serve themselves. Communicating
purchase requirements, checking order status, selling on-line,
and providing timely content are all good examples of this. In
addition, eBusiness can increase profits through increased revenue
generation. Common examples include direct product sales and
advertising revenue.
Overview of Web Traffic Analysis Software
The web traffic analysis market barely existed a short two years
ago. Today it is a multi-million dollar segment of the software
industry. Clearly, this is being driven by the growth of the
world wide web and the desire to know as much as possible about
visitors, through self-identification, registration, and web logs.
So what exactly is a web log, and where does this data come from?
When you browse the web, you are merely making an internet connection
with a remote computer (called a server) and asking for one or
more files to be sent to you via TCP/IP (Transmission Control
Protocol/Internet Protocol). This protocol spits the file data
into little pieces called "packets" and sends them to
your computer via the Internet. These packets are then reassembled
by your computer and displayed via your browser.
Behind every web site you visit is a web server, whose purpose
in life is to respond to you by locating and sending the requested
files to your computer. After each request, the web server logs
the results of the exchange in a "log file". A typical
log file contains information about which computer made the request,
for which file, and on which date. Additional information is
sometimes recorded as well including the browser type, the IP
address, error codes, the referring site, and the destination
site.
Log file analysis tools simply take this information and try to
make sense of it so that intelligent conclusions can be drawn.
Simple things like how many total files were requested can be
easily calculated and reported. By looking for multiple requests
from the same computer during the same timeframe, more complex
things can be calculated, like the number of total visitors and
visits that were made to a site. By adding other information,
like advertising information, ad impressions and click through
rates can also be calculated.
A more advanced data collection method (called packet sniffing)
has recently been introduced into the web traffic analysis market
that eliminates the need for log files entirely. This technology
gets its information directly from the TCP/IP packets that are
sent to and from the web server. The advantage here is twofold:
- Data is collected in real time, rather than being read in
from a log file after the fact.
- Companies with distributed web servers can easily and automatically
collect information in a centralized data warehouse for later
analysis.
Some Comments on Privacy
Once people realize that their behavior on the web can be tracked
and analyzed, we are often asked about privacy issues. It is
important to note that there is no way to obtain private demographic
information from web traffic analysis software alone. This would
include your name, address, phone number, email address, and credit
card number. In this regard, the web remains a completely anonymous
venue.
That said, when you visit a web site, your movements on the site
are completely public. This is useful information when reported
in aggregate, as it gives the webmaster direct feedback on popular
content. It also benefits the user by giving the company direct
feedback on how to improve the content and navigation of the site.
Many savvy marketers also try to collect demographic and psychographic
information about you through on-line surveys and forms. This
is also not inherently bad. It benefits the visitor by providing
the company with direct customer feedback, which results in better
products and web sites. This level of information is also typically
supplied on completely volunteer basis.
Most of the concerns we have over privacy involve what a company
does with the data, not what data they collect. Beware of companies
that are in the business of selling or supplying your data to
others. Responsible companies will:
- Let visitors know what data is being collected
- Let visitors know how this data will be used
- Allow visitors to refuse to supply the information
- Educate visitors on the benefits they derive from supplying
personal data.
What Users Want in Web Traffic Analysis
Software
With over 30,000 users, Marketwave is arguably the leader in this
emerging market segment and has a founding staff who is grounded
in both technology AND in marketing. We have performed extensive
customer research to try to better understand what companies are
looking for in web traffic analysis tools. Based on a recent
Marketwave survey received from over 3,000 users of web traffic
analysis software, here are the top features that are being asked
for:
Note that the top requests are for basic capabilities such as
pre-defined reports and easy to use software. This would imply
that the majority of customers would like an easy way to get a
quick overview of what is going on with the site.
As far as what customers are using traffic analysis software for,
we got the following top answers:
Again, getting a general overview of the site is by far the most
popular request, followed by information that allows webmasters
to better design the site. Web Marketing style sites are next
in popularity, focusing on ROI and measurement of marketing campaigns.
Revenue generation was requested by about 8% of the customers.
Overview of Web Traffic Analysis Market
Segments
Based on our experience to date and the data presented
above, we see three web traffic analysis market sub-segments,
roughly corresponding to the stage a company is in with its web
site development.
Customers in this segment are looking for overview information
about their web site. Tools in this segment report basic information
that is already contained in web server log files. They use calculations
and assumptions to create a maximum amount of log data relationships
for inclusion in reports.
This information is typically used for two purposes. The first
is site management. Webmasters and System Administrators who
are responsible for keeping the site up and running, often want
to know how much traffic they are getting, how many requests fail,
what kinds of errors are being generated, etc. The second purpose
is to internally justify a company's investment in the web and
build a case for moving the site on to the next stages. Traffic
analysis tools provide insight into the design of the next generation
site so the company can more directly respond to what visitors
have been most interested in. This is an early awakening of personalization.
Typical customers for Web log analysis have web sites at the brochureware
phase of development.
Customers in this segment want to increase reporting capabilities
to support more complex analysis needs - typically to support
return on investment calculations. This analysis involves gathering
data about the market and adding it to the analysis.
The type of data that is added includes advertising groups and
names, filters, and virtual server information. This data is
used to parse and filter log data, as well as report advanced
relationships of grouped and named data.
Product Managers and Marketers are typical candidates for tools
in this segment. They are typically managing sites at the customer
support and on-line transaction stages of development.
Rather than look at the web as its own island of information,
customers in this segment want to integrate with other databases
in the corporation such as customer databases. Resulting reports
not only use advanced relationships between log data, grouped
and named log data, but also draw from external databases and
spreadsheets. The highest level of customers in this segment
gather data directly from "the wire", eliminating log
files as the primary source of information.
Many Web Miners adapt their offers based on traffic and, more
importantly, create new products that match the results of their
analysis. For these people, web traffic analysis is gold!
Business owners and managers are typical candidates for these
tools. They typically have sites at the personalization and community
stages of development.
A Summary of the Web Traffic Analysis
Market Segments
Web Traffic Analysis Market Segment
| Typical Customer Needs
| Marketwave Products
|
Web Log Analysis
Customers in this market have spent money on their Web presence and require a supported product to justify their web investment via a general site popularity overview. This is typically the "Brochureware" crowd.
|
General Site Performance
General Site Overview
General Visitor Overview
|
Marketwave Hit List Standard
|
Web Marketing
This market sees the Web as a major part of their overall marketing mix and are looking to optimize their marketing dollars. This is typically the customer support and/or on-line transaction crowd.
|
ROI analysis
Detailed Visitor Analysis
Content Performance
Ad tracking
|
Marketwave Hit List Professional
|
Web Mining
This segment sees their site as a revenue generator and profit center. Companies interested in this segment require a solution capable of scaling to meet enterprise-wide tracking. This is typically the Personalization and/or Community crowd.
|
Fastest performance
Webmining capabilities
Detailed visitor behavior analysis
Flexible report output and delivery
Flexible data storage
Data warehousing
Complete automation
Detailed ROI
|
Marketwave Hit List Enterprise
Marketwave Hit List Live
|
Recommended Review Criteria for Traffic
Analysis Software
There is wide variation among the many features and
capabilities of today's web traffic analysis packages. It is
important to keep in mind which features/benefits are the most
important for each market segment. A product feature set should
match its intended users needs.
For simple log analysis, most users are interested
in a basic overview of their site and look for easy to use software
with useful pre-defined reports. More sophisticated sites focus
more on web marketing and web mining features. This report has
highlighted the most popular and used web traffic analysis features
in a categorized and prioritized format.
We recommend the following weighting based on different
user profiles:
| Criteria Weighting
| Log Analysis
| Web Marketing
| Web Mining
|
General Criteria
|
60%
|
40%
|
40%
|
| Basic Reporting
| 30%
| 20%
| 10%
|
| Web Marketing features
| 10%
| 30%
| 20%
|
Web Mining features
| 0%
| 10%
| 30%
|
| Total
| 100%
| 100%
| 100%
|
This section outlines the detailed features that
customers look for in sophisticated web traffic analysis software.
After giving a definition of each item, good review criteria
questions are given. This list is categorized, and then
sorted in priority order, based on our marketing research.
|
General Criteria
In this category, Ease of Use, Performance, Flexibility and Accuracy are especially important. For users who want to do simple log analysis, Ease of Use becomes tremendously important.
|
| Ease of Use
| Ease of use is one of the top requests from users of traffic analysis software. While this is a tough category to put quantitative numbers on, it is certainly one that you know when you see. Many vendors in the category focus on "techno-elegance" rather than ease of use. Good questions to ask are:
How hard is the software to install and get up and running?
How intuitive is it to use?
Is the software organization report centric, or log file centric?
Do you have to learn a bunch of technical jargon, or does it just seem to work?
Are the reports easy to understand and professional looking?
Can you easily figure out how to make the customizations you want?
|
| Performance
| The obvious thing to do to measure performance is to take one set of log files and measure the time it takes to get a useful report. While this is a good first level test, it does not really make for an "apples to apples" comparison among the traffic analysis packages available. Rarely can you find two reports that represent exactly the same data. More importantly many packages, Hit List included, perform the most time consuming tasks when running the very first report. This information is then stored in a database, and is available for future reporting needs. We recommend measuring performance in terms of both the first report AND subsequent reporting sessions, which more accurately represents how users will really use the software
How fast and flexibly does the software create a simple report?
How fast and flexibly can the software create additional reports?
How fast and flexibly can report modifications be made?
|
| Accuracy
| This is a huge issue that the industry continues to wrestle with. All software must make assumptions about the data contained in a log file, as there are many ambiguities. What constitutes a visit and a visitor and two examples of this. Look for software that is endorsed by independent Audit Bureau of Verification Services, Inc. (ABVS), BPA Interactive (BPAI) and the Internet Audit Bureau (IAB).
Do the reports adhere to auditing standards for advertising set by ABVS, BPAI, and IAB?
|
| Documentation and Online Help
| This category includes both written and on-line documentation. The ideal product never requires you to refer to written documentation, as everything you need to know would be either obvious or included in an easy to find help file.
How intuitive is the product in the first place?
Can you quickly find what you need in the dialog box first?
Does the product have context sensitive online help that is actually useful?
Is there up to date help available on the Web?
How is the written documentation? Is it too technical or not robust enough?
|
| Web-servers supported
| Clearly, it is important to support log file formats from the most popular web servers. In addition, many customers have multiple web server environments, which require products that can combine multiple formats into one reporting database.
Which web-servers are supported?
How are multiple servers from different companies handled?
Does the product support proxy server logs?
|
| Platforms
| Web analysis software does not have to run on the Web server. In fact, it is most common to run the reports on a separate machine, to avoid bogging down the server. Software that runs on common platforms, but supports multiple log file formats allows for the most flexibility.
What platforms are supported?
Does the software have to run on the Web server?
|
| Databases
| Products with standard databases are generally preferred over those with proprietary formats. This allows for the database to be extended as necessary.
Is the information stored in the database actually useful?
Can the database be opened and edited with an industry standard product like Microsoft Access and SQL Server?
|
| Extensibility
| The most sophisticated customers require software that can be extended and integrated into their environments. Features like having an open database and open architecture are important to this group.
Can the software be easily tuned to do what you need?
Can you easily extend it to meet your needs?
|
| Upgradability
| As Web sites grow and become more sophisticated, it is highly likely that customers may outgrow the software capabilities. Look for companies that have product lines that users can upgrade to.
How scalable is the product?
Can the product be upgraded to more sophisticated versions as users needs increase?
|
| Company
| One of the most important, but overlooked categories is that of the company itself. You clearly want to buy software from a company that is committed to the category for the long-term. If the company goes out of business, you lose your investment in the product, the databases, and all the time spent training users. Pick a company with a track record of success and ideally one that is economically viable (profitable).
What is the company's reputation?
What types of companies already use the product?
What do other users say? Are references available?
Have you tried calling their technical support line?
How are the products reviewed in the media?
|
| Basic Reporting Criteria
|
| Pre-defined reports
| This is the number one requested feature for Web traffic analysis software. Customers would like to see pre-defined reports that already do most of what they want. In general, customers don't want to spend lots of time defining new reports and dealing with reporting layout.
How many pre-defined reports are there?
Do they already contain the information you need?
|
| Report Customization
| In many cases, customers need to create their own custom reports. Users want software that is easy to understand and customize.
Is it easy to modify the colors, fonts, and layout of the reports?
Can every element in a report be customized? If so, how?
|
| Pre-Defined Report Elements
| For custom reports, look for packages that offer a toolkit of pre-defined reporting elements. This allows users to easily build complex reports from a wide variety of Web traffic analysis calculations.
What reporting elements are available to the user?
Are these elements customizable?
|
| Report Filtering
| Often, users want to look at only a subset of the web data. This lets users focus on just the relevant information, through focused reports that are easier to read.
Can reports be filtered? What specific information can be filtered upon?
How easy is it to add a filter?
Can filtering be performed on a global, report, and element level?
Does adding a filter require re-parsing the log files?
|
| Report Output
| Different users require different reporting output formats. Look for tools with multiple output formats, including formats that can be posted or emailed.
What output formats are available? (Word, Excel, HTML, ASCII, etc.)
Can one report be sent to multiple output formats at the same time or do you have to run reports multiple times?
Can reports be automatically distributed to a predefined e-mail list of recipients?
|
| Trend Reporting
| This allows users to easily compare statistics over time and see how they change. This is ideal for long-term forecasting, event analysis and overall Web traffic analysis.
How easy is it to show daily, weekly, and monthly trends?
How are reports that involve long term trends handled?
|
| Multiple Virtual Server Reporting
| This capability saves time by automatically running an individual report for each of multiple, virtual servers rather than running the same report, one at a time, for each virtual server. ISPs and customers who are responsible for multiple sites are the most likely candidates for this feature.
How are multiple servers handled?
How can individual reports be generated for each virtual server or user directory?
|
| Remote Reporting
| Often, people in remote locations, or people who travel, need reporting information. Software that allows reports to be securely generated and viewed from the road offer the best solution.
Can reports be generated remotely?
Does each remote user have to have software configured or does it work any Web browser?
How is security handled in this situation?
Can different users be given different levels of remote reporting access?
|
| Title Name Lookup
| To improve readability, users often want reports in terms of title pages, rather than cryptic URLs. Ideally, users have the flexibility to choose whether reports show the 'title page' of the Web page and/or the URL of the Web page in reports.
Can reports show URLs, Title names, or both?
|
| DNS Lookup
| IP addresses (contained in log files) are difficult to read and remember. Customers require reports in more human readable domain name format. Look for software that can converts IP addresses into an actual domain name (e.g. 206.129.192.10 = www.marketwave.com).
Does the product automatically perform reverse DNS lookups?
If so, how quickly does this run (it can take a long time if done serially)?
|
| Report Scheduler
| Many users want to automate the reporting task for periodic updates, or to run overnight. Look for packages that make this process easy. This is convenient for those who are not always in the office, yet want to run reports at a pre-determined time.
Can reports be scheduled and run automatically?
For Windows NT users, can the software be run as a service?
Can database maintenance also be scheduled?
|
| Web Marketing Criteria
|
| Advertising Reports
| For users who run on-line advertising, look for software that can track and compare results. This provides verifiable advertising information such as click-through rates, ad impressions and ratios. It helps Web marketers understand the quality of their ad banners by reporting visitor's path, from click through to exit page.
Can advertising effectiveness be reported?
Can on-line campaigns be easily compared?
Can entire advertising campaigns be grouped together in a logical fashion?
Can the product compare relative lead quality for visitors originating from each ad banner?
Does the product allow you to compare the effectiveness of running the same ad on different pages of a site or from different sites?
Are the ad statistics verified accurate by ABVS, BPAI, or IAB?
|
| Keyword Reporting
| This tells users what visitors were searching for in search engines like Yahoo to find your Web site. This is an effective feature for companies wishing to improve their promotion and presence on the popular search engines.
Is search keyword reporting possible?
If so, can you easily get a report for each individual search engine?
|
| Query String Parsing
| This capability allows Web marketers to understand what type of information their visitors are searching for in databases on their own site. It is popular to provide dynamic content where pages are served by a database. Query string parsing is valuable information for repositioning content, developing marketing events, etc. It is essential for database driven sites that use Microsoft ASP pages or Allaire Cold Fusion.
Is search parsing supported?
Can query string parsing be easily changed to work with any database?
Can the product detail what content was served and what the visitor typed in to find that content?
|
| Crosstabs
| Sophisticated web analysis users want to be able to analyze data by market segments, rather than simply as a whole. Cross-tabbing provides the ability to cross reference and compare one or more data elements - which is important if you want to compare visitors by country against which operating systems are most popular in each for example.
Are cross-tabs supported? If so, how easy is this to set up?
|
| Grouping
| Many users want higher level reports, using groups, so that related information can be combined. This puts information into more meaningful and readable top-level groups. Typical examples include showing all visitors from AOL, CompuServe, Prodigy and MSN in a single group called Online Services.
Can reporting data be grouped into higher level categories?
How easy is it to manage these groups?
How easy is it to push down and get more detailed group data?
|
| Cookie Reporting
| Cookies are used to better track individual visitors, especially those that visit multiple times. For sites that use cookies, web analysis software should take this info account and individually analyze which Web sites your visitors came from and which paths they took through your site.
Are cookies supported? Do you have to use a proprietary cookie DLL or can you use any cookie system?
Are both session and persistent cookies supported?
If so, can a path analysis be done based in this information?
|
| Web Mining Criteria
|
| "On the wire" data collection
| Many high-end enterprises with sites in multiple locations prefer to eliminate the need for log files entirely. Look for packages that can read network packets right from the wire to handle this.
Is live data collection supported?
If so, are log files supported as a backup?
Is data collection limited to specific web server/operating system combinations?
Does the data collection process have to be installed on the web server or can it be installed on a separate machine?
What is the performance hit, if any, to the web server?
|
| Data Distribution and Data Warehouse Support
| This allows data to be stored locally, as well as transmitted to a centralized enterprise wide database. This is useful to large organizations with distributed sites around the world.
Can Web traffic data be simultaneously stored locally and to a remote global data warehouse?
Can the product store information coming from multiple data collectors from all over the world?
Can the product compensate for different time zones?
|
| Data linking
| Many enterprises already have data on their customers, partners, suppliers, etc that they would like to combine with the data from their Web site. Look for software that allows web data to become an integral part of the enterprise, rather than yet another island of information within it.
Can data from existing databases be incorporated into reports? If so how?
Is this supported via an easy point and click interface or does it require sophisticated database programs?
|
| Database Manager
| High-end sites will quickly create multiple gigabytes of log data. They will need to routinely purge old data from the database and/or selectively import existing data into the database. In addition they need an easy way to delete existing databases, condense databases and move between multiple databases.
How are large amounts of data handled?
What database management tools are included?
|
| ODBC Database Support
| Logging to an ODBC database allows for the storage of an unlimited amount of data without having to maintain multiple databases for analysis.
What high-end ODBC databases are supported? (SQL?)
|
|
