October 2006
Understanding Web Traffic: A Three-Part Guide for Business Owners
The Problem
The majority of small business owners don't take advantage of web traffic data. After all, it's often difficult enough to develop a website that works. Compound the usual demands of small business with the technical language of web traffic, and web statistics become something that's easy to set aside. Technical people often aren't asked to create web traffic reports. Most small businesses don't have the luxury of a large staff to dedicate a position to web analytics, a role which can bridge the gap between business needs and technical know-how. So rather than traffic reports, we reply on customer reports: "I love the new site," "I had a hard time finding the right product," "I didn't buy online, because you don't accept American Express."
Why Bother?
Imagine for a moment the following scenario: at a networking event you're talking with a retail store owner, learning a bit more about his business. You ask how people find out about his shop. "I don't know," he answers, "I just keep my eye on the register." Sales do keep the wheels moving, you agree. Where does he market his store? "I don't do much marketing... the phone book, I guess." How have his customers learned about him? "I'm not sure. Word of mouth?" From whom? "I really don't know," he answers. Well, how many customers come into his shop? "I don't track that. Really. All I do is watch my sales."
We'd probably think this store owner peculiar; maybe just lucky to have a self-running business. While sales are the lifeblood of any company, we know intuitively they don't provide the sole picture. To the contrary, most retailers closely track coupon effectiveness, marketing campaigns, foot traffic, product placement... why not bring this level of sophistication to the web as well?
Well, it is possible for a small business owner to move from "I think people look at my web page" to "8% of the people who view our revised home page target message go on to buy our products." In this three-part series, we'll cover the basic groundwork and vocabulary of web traffic, discuss how to create useful reports for typical business questions, and review some of the software tools that are available to help make sense of it all.
The Groundwork
Web statistics work in a surprisingly simple way. Every time a user views a page, their web browser talks to the web server by requesting a specific "page" through the URL -- the Universal Resource Locator. The URL often ends with the name of a file ("about_us.html"), which the server fetches from storage and sends to the browser. That file, in turn, might include a business logo, pictures of the president or the board, page graphics, a typographic style sheet.. all sorts of things that go on to make up the page as it loads. It's the server's job to send each of those to the browser as well. On a slow dial-up connection you can literally watch the page be constructed before your eyes, element by element.
Each of these requests stores a bit of information to the server as well. Obviously, the requested file. But also the date and time, possibly the web page that the user was at beforehand, and some information about the web browser that made the request. This might include the browser type, the type of computer, and the computer's IP address -- a unique series of numbers that identifies every computer on the Internet. These data are the building blocks for web statistics. Everything is built from here.
Key Terms
Page views are easy enough: a count over a period of time of the number of times a web page was viewed. In many ways, this is the most useful and important count of your web traffic. If normally 100 people view your home page on a given day, and suddenly this jumps to 800 for about a week and falls off, then you know something interesting has happened. Was your company mentioned in a newspaper article? Did a newsletter campaign link to the site? Have bloggers found something interesting and linked in? All of these might be reasons.
Sessions are a bit more abstract: a session is composed of a user browsing a set of pages over a given period of time. So, if one user clicks through to your home page, your about page, and then the contact page, then picks up the phone to call, that would count as three page views and one session.
How much time? Here's where some problems begin. Web traffic reports are always just estimates. Typically, we assume that a session consists of a user clicking around a site, where each page is within 20 minutes of each other. If the user goes to lunch, comes back, and continues surfing, that would count as a new "session."
And how do we identify the user? Again, another problem. Web statistics can't actually track users -- they track IP addresses, numbers assigned to computers when they're connected to the Internet. And these numbers can change. Some DSL suppliers, for example, change everyone's IP address once an hour. Other users have "fixed" IP addresses. So again, sessions count 'users,' but this is really an approximation.
Hits are, unfortunately, the most over-used, least understood, and least useful measure of web traffic. This is, in part, because it's so fun to say things like, "our website receive 50,000 hits this past month!" But in the above example, each element that was downloaded to create the about us page would count as a hit: one for the page, one for the logo, one for the president's headshot, a dozen or so for page graphics, etc. Since the number of elements that creates a web page is arbitrary, hits don't count much that's useful for business reports. Ignore them, and focus on page views.
This Month's Action Items
To get your own business started on understanding web statistics, try one of the following:
- Have bi-monthly web statistics reports emailed to you. Your web provider or office guru should be able to configure your web statistics software to email reports. Have a basic, "executive view" report sent to you twice a month. Include the most popular pages of your site, your overall page view traffic for the time period, and (coming in the next installment) the site that sends the most referrals to you.
- Evaluate your site content from your customer's perspective. As a user, what pages do you think would be most important? Create a short-list of your key site pages. Then, look at your traffic: do the feet in the door reflect your priorities? Why or why not?
- Look for surprises or irregularities. Is a particular product page getting far more traffic than you expected? Perhaps it should be highlighted on the home page! Are people searching for particular content? Make it easier to find! People will remember your site as being helpful and friendly if they're able to better use your site.
- Evaluate your traffic before and 4 weeks after a promotion or campaign. Are particular pages more popular than before? Are products being moved as expected? If your campaign was an HTML email, you should see an uptick on the newsletter's landing pages, as well as associated pages. Find out what people find interesting about your site!
Next Steps
Even with this simple introduction, as a business owner you can now move from the meaningless ("50,000 hits last month!") to the useful ("Last week, our home page page views increased by 30%."). A great start! But there's much more that can be done. Consider:
- How valuable is an advertisement on an industry-related website?
- Of our affiliates, who refers the largest number of customers?
- How much income has some particular advertisement generated?
- How many people fail to complete the checkout process, and where do they abandon?
To answer these questions, stay tuned for parts two and three of this article. We'll explore web traffic tools, identify site goals, and learn about tuning up one's site to obtain the best results from these analytics.
