 For most sites, the info stored in cookies can be a rich source of data that can be mined in many different ways. And if you actually write cookies for the purpose of data analysis, you can extract practically any information you desire from your site.
It's likely that you have much more about data your site visitors than you may think…you just have to know where to look, and what to do with what you find. In this article, I ’ll provide seven tips for mining data with cookies. From determining browsers and resolution to tying in member status or role, these tips represent the proverbial ‘tip of the iceberg’ when it comes to using cookies for effective analytics. Take our tips below as inspiration for developing cookie sets that provide exactly the information you’re after.
1) Determine what web browsers are being used
ClickTracks intentionally avoids overly-technical reports such as what web browsers (user agent) visitors are using. Why? Because we found that most marketing folks don't give a hoot about knowing if someone's using IE 6.0 or 6.1 SR-2. However, if you're an atypical marketer (and we mean that in the nicest way possible, of course), and must know what browser, the solution to the �browser question' is to simply apply a common snippet of code to your web pages.
The first part of the code finds out what the users browser is; the second piece writes that information into a cookie. This code can scripted in JavaScript or any of the server-side scripting languages such as ASP, JSP, PHP, etc. They all have the ability to determine the user's browser.
If you want to take it a little further, you can use your scripting language to modify the user agent strings, perhaps trimming them down into just three categories such as IE, Mozilla, and others. Or maybe you just want to see how many visitors are using �antique' browsers, so you separate them into two categories: old browsers and new browsers. Suffice it to say that there are a number of ways you could massage this data with your scripting language before dropping it into the cookie, making data tracking in ClickTracks as easy as creating a label to track that specific cookie.
2) Determine what monitor resolution your visitors have
Does your information fit in the default window for the majority of your visitors? Using JavaScript you can determine what monitor resolution your visitors have. Perhaps you want to determine how many of your visitors have 600 X 800 pixel monitors vs. 768 X 1024 vs. 1200 X 1600, and so. A piece of JavaScript placed in the <head></head> of your pages can determine what resolution their browser is using, then write that value into a cookie. Then in ClickTracks' Advanced Labels menu, you'd select the resolution cookie, and could see all the various browser resolutions in the list. (You might be surprised just how many different browser resolutions there are!). And much like the web browser example, you may want to consider massaging that data before writing the cookie�possibly breaking visitors down into three groups like low, medium and high resolution.
3) Determine the operating system visitors have
Does your company sell products targeted toward people who use Macintosh or Linux computers, rather than Microsoft machines? If so, you may be interested in knowing what operating system a visitor to your site is using. Like monitor resolution, a visitor's operating system is another piece of information that your scripting language can determine. Dropping OS info in a cookie and creating a label based on operating system (e.g. one for Linux, one for OS X, etc.) means you'll be able to see how long those people stayed on the site, (from the Site Overview report), what pages they were more interested in (from the Navigation report), what key words they responded to (from the Search report), and all of the other information from the various reports.
4) Time since last visit
If you drop a cookie that includes a timestamp of when they visited the site, you could check for this cookie on subsequent visits, determine the time between visits, and drop that information into a cookie. Of course, you wouldn't want to just drop the exact time difference in. If you did that you'd get as many different values in this field as there were unique visitors. Instead, you would want to do some manipulation to establish a few discreet groups, such as: less than 1 hour, 1 hour to 8 hours, 8 hours to 1 day, 1 day to 3 days, 4 days to 1 week, 1 week to 1 month, etc. This way you'll have just a few values for this cookie, and you can easily create meaningful labels around them.
5) Number of visits (i.e. frequent or infrequent visitor)
In a similar vein to the previous item, you could also create a label that increments each time a visitor is recognized as a returning visitor. So within that cookie you would have a count of the number of visits to your site for each visitor. Again, in order to create meaningful labels you may want to do some grouping and create a cookie that has discreet groups such as 1 to 5 visits, 6 to 10 visits, 10 to 20 visits, and so on. The labeling system in ClickTracks does not allow you to do mathematical operations. So you couldn't create a label for everyone with more than 1 and less than 6 visits, for example. Therefore, any mathematical operations necessary to establish these groups should be done with your scripting language on the site itself. Then you simply select the values from a manageable list in the label wizard, and you'll quickly be able to tell which pages your frequent visitors really like.
6) Geographical labels based on state, region, zip code, etc.
Perhaps your business is dependent on locality, and you want to know where people are coming from. The server or client environment doesn't contain this type of information, so to create cookies for this you'll need to gather data from a database. If you have a back-end database storing information about registered users, you probably have data there about where their home or office is located. And assuming you can identify your registered visitors by way of a cookie, you can then retrieve locale information and drop that into a cookie. Again, you may want to do some manipulation of the data before dropping it in the cookie. Instead of dropping state information, you may consolidate states into regions. Instead of dropping entire zip codes, you may just record the first two digits of the zip code. Through your scripting language, it is within your control to store data in whatever form is the most meaningful to your business.
7) Member status or role
If you have a backend-database of registered users, they may have different roles or status. For example, you may have regular members and premium members. No doubt this information can be retrieved from your database and placed in a cookie. This will allow you to create labels that will tell you how your premium members behave, relative to your standard visitors. Of course, this is just one example. Taken further, you can do additional queries or scripting on the web server to determine if a member has spent a lot of money on your site, or a little, and thus compare your big spenders vs. small spenders. Any way that you can identify a type of member from data stored in your database can be recorded in a cookie and labeled in ClickTracks.
Hopefully this article has expanded your thinking about the power and flexibility of data mining through labeling cookies. Certainly the process requires some thought and web development efforts. But there is a wealth of knowledge available that can render returns far beyond the cost to set it up.
 |
Subscribe to this comment's feed
Visit EmailLabs 
