Statistics on the web: What they represent - Part 1

So you have heard the words hits and visits a million times and you still are not quite sure what they mean? You will notice we are very proud with ours and we show them every month on the top-left corner of our home pages. This article aims to explain some basics about statistics on the web, what do they represent and how to gather them. First of all, let's start with a few basic definitions. A "visitor" is a count for every person that has visited a site and has spend some time on the site (the duration of the stay is known as a session). Visitor counts depend heavily on how a session is defined (a set amount of time or some other parameter). A visitor count may also be referred to as a "session" count. Some statistics may give you a "unique visitor" count, which represents the number of individuals or computers that have been to a site for a given period. This counter is similar to the visitor counter, but it ignores repeat visits (sessions) for the given period. There are flaws in how individuals or computers are uniquely identified (by their IP addresses or cookies), so this counter is not going to be 100% accurate. The most reliable and common statistic is "page views," also referred to as "page hit." It measures the number of times a web page has been seen by users. The last (most granular) counter is a "hit." It is not to be confused with a "page hit." A "hit" is registered when any element is being accessed on a site (not just webpages). An element can be a webpage as well as an image, javascript or a number of other support files. Some of the elements don't have any meaning to a human - they just help the browsers display pages at their best. Hits is becoming largely a meaningless statistics, as it can easily be abused. Pages can be designed (on purpose or not) to force a large number of hits for small number of visits. Some modern techniques (like AJAX) also tend to increase the amount of hits on a server. Additionally, hits can vary widely from one period to another, due to changes in content or web design and should not be considered as a good indication of increase or decrease in traffic. Hits are a more meaningful statistic for measuring server load rather than human traffic to a particular site. Additionally, some useful statistics you would want to look into are: *OS/browser/resolution your clients are using, so you make sure your site looks best at the most common combinations. *Geographical locations, so you can target local information to your visitors. *Percentage of new vs returning visitors and time they stay on site (pages and minutes) to give you a better idea of how "sticky" your site is. Do people find content they like reading and getting back for? *Traffic sources - understand where your people come from. Are they mostly coming from search engines, links from other sites or direct traffic (email/bookmarks/offline advertising)? What keywords they type in to find your site? Armed with this knowledge about website statistics, next month we are going to cover how to collect them and what tools can help you with that task. Ivan Peevski is the chief technical officer of the New England and New York Real Estate Journals, Accord, Mass.