What's the difference between the free stats programs offered by most web hosts and professional analytics applications like WebTrends 7.5 professional?
Most people are trained to think of their web stats as nothing more than a way to track unique visitors and page views. For a low traffic site, or one that is no more than a forgotten brochure for the business, the difference is perhaps unimportant. However, for anyone considering their web site as a valuable part of their marketing toolset - the difference is enormous.
Data without reference points is meaningless
Tag based analytics enables the marketer to provide meta data about the information on the web site so that the analytics software can provide relevant and interesting
information on the traffic and usage of the site.
A web server cannot tell one page from another from the eyes of a visitor or marketer. The un-tagged information on a free web stats program will not show if certain kinds of content are more popular than others. It will not show you organic search traffic from paid search, which types of visitors are more valuable, most popular paths through the site, or which groups of pages are more useful than others. To extract this kind of meaningful data about your site the marketer needs to teach the software about the site and its traffic.
Information rich data becomes action-able knowledge
The uses of tag based are as varied as the types of sites it can report on. Lead generating sites, e-commerce sites, advertising supported sites, customer service sites, portals and more can all use tag based analytics to provide business intelligence. It enable marketers to consistently improve their results in iteration after iteration of their web site changes. It enables them to target advertising more effectively. They can improve conversion rates, improve content effectiveness, improve user experience, and more.
Accuracy of Information
Before I move on to give specific examples of the information tag based web analytics can provide I think it is important to discuss an issue of accuracy provided by something called client side tagging versus log based statistics:
All free or low cost statistics programs, (and even higher end analytics software unless configured otherwise) will generate statistical data based on the log files created by the web server.
This has a serious flaw.
Web server log files are the raw data collected by the web server for every hit by a visitor. Every time s request is made on the web server, a hit is recorded as a new line in a text file. This can be a request for a page, an image, a PDF, an included object, or any other file on the web site. Each hit is recorded with a timestamp, and the IP address of the computer requesting it, the browser used, and other various and configurable information.
The statistics program simply reads these log files and organizes the information into human readable reports and graphs. Each unique IP address that shows up in the log file is considered a unique visitor.
The problem is, s unique IP address is not enough information to differentiate one user from another. Five users behind a DSL line in a small office will look like one unique visitor. One hundred users in an medium office may look like one visitor. Tens of thousands of AOL users all over America will look like one user. One user from Herndon, Virginia, in fact - because AOL uses a massive proxy server there so it can cache the most popular content on the intent to deliver it more quickly.
There are other log file methods, such as combining IP address and computer type to differentiate visitors, but the limited ways a standard web server can differentiate between one visitor and another is limited and ineffective. Because of this problem log based statistics should be considered by all to be inaccurate.
To solve this problem tag based analytics software uses a technology called client side tagging to uniquely identify each computer visiting it's website - regardless of whether they are in an environment that would otherwise fool the web server into counting them incorrectly. This "client side tag" is a few lines of JavaScript on each web page being analyzed by the analytics software. Each time one of these pages is requested, the computer requesting it automatically runs a script on the page, and reports back to the analytics server to identify itself as a unique computer.
The result is significantly more accurate statistics (nearly 100%), that typically show a large increase in the number of visitors being counted.
Tagging for content usage
Now that I have explained the accuracy of information issue in the two types of reporting, lets move on to the more exciting benefits of tag based analytics.
Paths and Pages
Starting from the most basic level, page name tagging allows you to provide meaningful names to the pages on the site so that it is easier to tell which pages are being visited.
Without tagging, the only ways to view a page in a report is either by URL (address of the page), which can be long, cryptic or meaningless, or the HTML title - which for un-optimized sites are often all the same, and for search optimized sites are often long and difficult to differentiate.
When you teach the analytics server what the pages mean to you - the results are just easier to read. Home for home, Products for products, etc. When you have hundreds or thousands of pages on a site, especially similar ones, this makes a big difference.
Web analytics software can also measure the order of pages visited by each visitor. It will then calculate the most common paths through the site. For example, an e-commerce sites most popular path may be - home, products, product a, product b, search, exit site.
This kind of information can be enlightening. For example, why don't most visitors go to product c? Why don't they add anything to their shopping cart? If everyone is searching what's wrong with the navigation? Is there something wrong with the search results? You can't test and measure without this kind of information.
Content groups and sub-content groups
On my own site, I have four major categories - email marketing, search engine marketing, web analytics, and reviews. Each category has many pages. Since I have tagged all of the pages in each category as a group, I can look at my online WebTrends reports and see which type of content is more useful or more popular with my visitors. This can be an indication of what the market is most interested in. I can check it against my sales and notice correlations or discrepancies.
To make it more interesting, each major section on my web site is broken down into products, services, and knowledge (frequently ask questions). I have tagged these as well so I can see which sub-sections are most popular.
Pages can also have more than one content group and sub-content group. This allows you to have pages show up in multiple group reports - if product A fits under category A and category B, no problem - just tag it correctly and you will see it in both category reports.
To continue with my example, I have tagged each page on my site to have two groups and sub-groups - in reverse of each other. So now I can look at the sub-sections as major sections to see which major sections have the most popular sub-sections.
Basically, any way you want to group your content can be achieved through a well planned tagging strategy.
Tagging for scenarios and funnels
Tagging for conversion is cool. Let's say you have a membership based site and you want to track how many visitors follow a certain path on your site. For example, you may want to track what percentage come to the home page, click the "learn more" page, click the "sign up" page, and then complete the sign up process. No problem.
In WebTrends, you would tag the series of pages as a scenario. You may choose to call the Scenario "perfect user", or "new signup", etc. You would tag each page as scenario step one, step two, etc., and name each of the steps. The end result is a beautiful graphical report that shows you how many visitors started each step, and what percentage went on to the next step.
Additional, invaluable information in this report is the "leakage" in the steps. Where do visitors go from each step besides where you want them to go? This is the information you need to make improvements on your site that will improve your conversion rates.
Tagging for conversion and sales data
For e-commerce and lead generating sites tags give you the information you need to improve your marketing spend. With tagging you can pull sales data into your WebTrends reports. WebTrends can pick up the items purchased and total sales from your 'thank you' page and display this information in your reports.
For search engine marketing (paid and natural search) you can break down your sales activity by keyword phrase and search engine. You can also see it by e-mail campaign, links form other sites, type of visitor and more.
Geographical reporting
It's one thing to see your visitors broken down by country, but what about province/state, city, area code, company, or designated marketing area? All possible thanks to the built in, up-to-date geo-IP database called GeoTrends. The GeoTrends database knows the location of every IP address in your reports and can therefore provide detailed and accurate visitor location information.
Getting the most from web analytics
A professional analytics program like WebTrends essentially allows you to track whatever data you need and turn it into action-able business intelligence. The functionality for this exists out-of-the-box, however to generate the most interesting reports for your business requires some effort on your part.
The ability to tag information on a web site comes with a responsibility to think through the data on the site and what kinds of reports need to be generated from the data collected.
At Agito, we go through a detailed process on every web site we track to ensure that we have identified all of the information we need and tag all of the pages properly. We go through a mapping process to define all of the desired paths through the site. We consider email campaigns and advertising campaigns and plan landing pages that can be tracked. We build spreadsheets of the web site to list every page and the tags that need to be associated with every page to achieve the desired outcomes.
If you want the most out of your web analytics, make sure you go through a process that will uncover your needs and then teach your software to deliver the results you want. You will be way ahead of those who use only free tools, or even professional tools with only out-of-the-box functionality.
