What's Your Web Site Up To?
Here are two programs to help you analyze those log files that IIS uses to audit Web site activity.
If you’re using Internet Information
Server to handle your company’s Web site, you’ve
probably noticed that it saves log files. Perhaps you’ve
even opened one of these log files up in a text editor and
realized that it has complete information on who’s visited
your Web server and what they looked at when they visited.
But like any other mass of completely raw data, Web server
log files are difficult to extract useful information from.
Fortunately, there are applications designed to help you with
this task. In this review, I’ll look at two programs:
Analog and Statistics Server.
Before you start analyzing Web logs, you
need to understand a little bit about them. Web logs capture
the HTTP requests that come into your Web server. These requests
include a lot of information, including time and date, name
of the browser, requesting IP address, and requested page.
There are several standard formats for Web logs, including
one Microsoft developed that IIS defaults to using, and another
from the W3C that uses an extended log format. I recommend
the W3C’s extended format to capture the most information;
you can make the change in the Root Web Properties dialog
box within Internet Service Manager.
Once you’ve got a few days’ worth
of logs, the easiest way to look for patterns is to use Analog,
a freeware tool. Analog is available as source code or as
an executable for a wide variety of platforms, including all
versions of Windows that can run IIS. Analog’s operation
is simple. First, create a text file that tells it what reports
you want. Second, run the program. Third, open the output
in any Web browser (see Figure 1).
Figure 1. Analog gives you a quick view of your
IIS Web logs; this is a sample of one of its daily reports.
A few of the reports I find useful in Analog:
Number of hits by hour, day, week,
Which pages are the most-viewed, by
number of views
Which pages contribute the most load
to the server, by number of bytes downloaded
What search terms people are using
to find the site
What other sites are sending business
Which countries visitors come from
What pages failed to download
Surveys have shown that Analog is the most
popular log file analyzer around, and it’s easy to see
why. The program is very fast, it produces sensible output,
and it’s free. Another measure of its popularity is that
people have written Analog add-ins to handle configuration
chores or create pretty graphs from the output. If you’re
running a Web site at all, you really ought to try this program.
Sometimes getting reports after the fact
just isn’t enough, though. In that case, you’ll
want to take a look at a program such as Statistics Server
Live Stats 5.02, from MediaHouse Software. Statistics Server
still relies on Web logs generated by your server, but it
reads the log file continuously (you can set the interval;
it defaults to once every two minutes). Statistics Server
software then translates the log file entries into a series
of Web pages that you can access from its own server port.
Configuration of Statistics Server is handled
through the same browser interface that displays the reports.
This includes choosing which server to analyze, setting security
for the Statistics Server, and so on. Help pages in HTML format
walk you through the process of setting up for the first time,
and you can be looking at data within half an hour of downloading
Statistics Server automatically creates
a variety of useful pages, including:
A summary of who’s logged on right
now, and what pages they’ve been looking at
Monitoring of particular pages or IP
Summaries of traffic by day, week,
- Most popular pages
New referrers in the last week
That’s just a sample. The bottom line
is that Statistics Server lets you keep an eye on your Web
traffic as it’s happening. You can use this just to reassure
yourself that everything is working correctly, to watch for
failing links, or to track the progress of an online promotion.
It’s a well-designed program with lots of functionality.
Figure 2. Statistics Server Live Stats produces
real-time, configurable reports of your Web site's visitorship.
You can download a fully-functional 30-day
trial from the Mediahouse Web site and see for yourself. Pricing
for the registered version starts at $395 for a version capable
of tracking the activity of up to 50 Web servers.
If you’re running a Web site without
some sort of Web log analysis tool, you’re running blind.
With Analog being freeware, there’s no excuse for this.
If you’re doing any sort of e-commerce or other marketing
activity with your Web site, you should consider how much
up-to-the-minute information is worth to you and evaluate
Statistics Server Live Stats as well.
Mike Gunderloy, MCSE, MCSD, MCDBA, is a former MCP columnist and the author of numerous development books.