measuring traffic
| The first step toward understanding your site and therefore evolving and improving it is measuring your traffic. Measuring traffic is as important as counting money to a web business. Unfortunately, it's a little harder. There are several different approaches to counting and tracking users, and I'm sorry to say none of them is exactly accurate. It's always best to decide how you're going to measure traffic before you launch your site partially because your decision may affect site structure and partially because you'll want to have traffic reports from the beginning. Collecting these initial reports allows you to do a full comparison as your site grows (it's not usually possible to do retroactive reporting). If your site is already launched, don't wait for a redesign to put a measurement system in place. Choose a system, and begin running reports as soon as you can preferably on an ordinary month, so you can establish a baseline comparison for any future changes. how traffic data is collectedTo explain how traffic is measured and analyzed, we need to first review some basics: When you visit a web site, your computer sends a message to the site's server, requesting the page and any accompanying graphics or scripts that appear on it. This message, or request, is usually called a "hit."
Traffic measurement is inexact. One visitor may be counted as 10, and 10 may be counted as one. When you visit a web site, the site's server log (think of it as a guest book), will usually record some data about you, or rather, about your computer: the computer platform and browser you're using, the domain you're accessing from (such as earthlink.net or berkeley.edu) and most importantly the IP address that uniquely represents your computer. (Note: IP addresses are temporary identifications assigned to each user during each session; they don't stay with you over time.)
In addition to the IP address, the server may place or look for a "cookie" on your computer. A cookie is a cute code name for a unique tag, which a site may leave with your browser, so it can identify you when you return. The IP address and/or cookie will help the site's server track you through the web site. By the time you move on to the next site, the server log will have documented your visit, including the number and order of pages viewed. The server log, however, contains only raw data. This data must be analyzed with software (like WebTrends' Log Analyzer) in order to return meaningful statistics, like unique visitors or pageviews per visit. (See choosing software for traffic analysis, p. 247.) an inexact scienceIt's important to realize that traffic measurement (like vote-counting in U.S. elections, apparently) is not an exact science. You can never know exactly how many users come to your site. No one can. The sites with the most accurate numbers are those that require users to log in before entering the site or using key areas. But this kind of tracking isn't feasible for most sites; they'd lose their users if they required them to log in. So the best bet for the rest of us is to count the individual computers as opposed to users that log on to the site. The limitations of this approach are obvious: If one person logs on from 10 different computers over the course of the month, she'll be counted as 10 different visitors. Conversely, if 10 people access your site from a single computer (in a public library, for example), they'll be counted as a single visitor. The uncertainty doesn't stop there. One of the basic and vexing questions (for sites that don't require log-ins) is how to track users as they make their way through the site: Are you receiving multiple hits from the same user, or single hits from multiple users? This is the question the server log asks thousands of times each day. Because there are so many ways to count and track users, no two systems will turn up exactly the same numbers for the same site. That's why it's important to choose one system and stick with it. The differences between reporting systems also make it difficult to compare numbers between companies. It's very difficult to ensure an apples-to-apples comparison. So your best best is to use the estimates reported by industry analysts Media Metrix or Nielsen Net Ratings. They both use a consistent methodology on all participating sites, ensuring a level playing field. Bear in mind, however, that the ratings will inevitably differ from your in-house measurements.
Learning the Lingo "Hits" vs "Pageviews" A pageview is a single viewing of a single page on your site, by a single visitor. A hit is a single request made for a single item on a page. To view a page, a user's computer must fire off many "hits" to your server one for the page, one for each image, one for each ad, and so on. So sites log far more hits than pageviews. Nonetheless, the two terms have become blurred. People often say "hit" when they mean "pageview." And that's okay in casual conversation. But if you're structuring a business deal, be clear about which term you're using, and what you mean by it.
|
EAN: 2147483647
Pages: 195