Article Index
5 Steps to Understanding HTML
7 Reasons Why You Need a Website
The Top 10 Biggest Web Design Mistakes
The Web Designers Toolbox
The Web is Not Paper
Theres More than One Web Browser
Time for User Testing
Titles and Headlines Its Not a Newspaper
Tracking Your Visitors
Understanding Web Jargon
Uploading Your Website with FTP
Using Flash Sensibly
Using Quizzes and Games to Get Traffic
VBScript Javascript Made Easy
Websites and Weblogs Whats the Difference
What Do You Want Your Website to Do
What You See Isnt Always What You Get
Which Database is Right for You
Why Doing It Yourself is Best
Why Java Will Drive Your Visitors Away
Why Word is Bad for the Web
Why You Should Put Your Content in a Weblog Format
Why You Should Stick to Design Conventions
Working With Templates
Writing for the Web
10 Easy Ways to Promote Your Website



Tracking Your Visitors

Tracking Your Visitors.

Once you've got some visitors, the chances are that you want to know more about them. How many are there? Where are they from? What web browser do they use? Luckily for you, there are plenty of ways to find out.

Server Log Analysis.

Most web servers keep a log of every file they send, with information about the request they received. These request headers contain all the information a user's web browser sends to the server when it asks for pages, images or other files. The information includes the user's IP address, their web browser's name and version, and the kind of files their browser can handle.

Out of this information, the IP address is probably the most useful. Each block of IP addresses is allocated to a certain ISP in a certain country, meaning that you can use them to tell roughly where people are from. There are plenty of free databases out there that map IP address to physical location, letting you break down your visitors by country or even, in many cases, by state.

The other thing IP addresses do for you is let you identify how many unique visitors you have – that is, how many actual people saw your site as opposed to how many pages were loaded overall. This lets you figure out things like the average number of pages each visitor sees, or the number of times the same visitor comes back.

You can get software that will take this information from your server logs and turn it into easy to view tables and graphs – in fact, most web hosts will have already installed some software like this, if you look under the 'statistics' section in your hosting control panel.

Cookies.

IP addresses can be influenced by all sorts of things, notably ISP proxies making a whole ISP full of visitors look like just one. As well as crude IP address tracking to find unique visitors, then, you might also consider using a cookie. All you do is leave a cookie on each users' computer with a randomly-generated ID number, and then check each visitor for cookies to see if they've been to your site before.

If you log how many ID numbers you give out and how times each ID number appears in return visitors' cookies, you can get a better idea of just how many visitors there were overall, and how many times each one came back. You should consider, however, that many users have cookies turned off in their browser, or ask their browser to prompt them to accept or decline each cookie individually, so while they're generally more reliable than IPs alone you can't depend on them completely. A mixture of the two methods is best.

Registration.

If you want to know more detailed information about your visitors, you can ask them to register and log in to use your website. This gives you an opportunity to collect their email address, their exact location, and pretty much anything else you dare to ask.

You have to understand, though, that many people will be unwilling to associate detailed demographic information with their identity. Also, it's difficult to get registration right: ask for it too early and people will just leave without seeing what you've got to offer, ask too late and they've already got what they came for.

Surveys.

As an alternative to registration, you might try including random surveys. This is the technique favoured by most big companies: simply pop-up some kind of message saying 'would you be willing to participate in a survey to help us improve our website?', and then pop up the survey questions if the visitor says yes.

The advantage of this is that surveys can clearly state that they're completely anonymous: you don't know the person's name, where they live, or anything else like that. This gives you the opportunity to ask more personal questions that you would otherwise be able to, establishing a solid demographic and preference profile for different parts of your audience.


Google

News Online