But there are a lot of misconceptions about the terminology that is used when people talk about Web traffic statistics. I am therefore going to:

• Discuss the terminology
• Make some suggestions as to which Web Stats package to use for web site analysis
• Make some suggestions on how to utilize your web statistics to gain a better understanding of how your website work (or doesn't work)

Hits, Pages, Visitors - what do they all mean?

A couple of years ago it was very fashionable to have a 'Hit Counter' on your website. It supposedly showed the whole world how many 'hits' you've received to date. As with all trendy fashions, the Hit Counter as we knew it has seen its day. Today you run the risk of being labelled a newbie, or a hobby website owner if you have a hit counter on your website. The problem with hit counters are that

1. Nobody is quite sure what these 'hits' are counting. As you will see, collecting detail about website statistics is not as straightforward as it sounds. A hit could be anything from a unique visitor to the fact that you've displayed 5 graphical elements on your site (in which case, 5 hits will be recorded!)
2. Hit counters, unlike odometers, are very easy to manipulate. It is not unknown for a desperate webmaster or website owner to set the hit counter to start at a vastly inflated number. Hit counters can also easily be reset - again to any number that you feel like.
3. If you have a website with very low traffic, do you really want to proclaim that to the world? Especially if you are trying to promote your website for advertising purposes (let this be a warning to anyone wanting to advertise on a website - first do some investigation into how much traffic the website gets before laying out big money).

As I mentioned, collecting accurate traffic statistics is not straight forward. Different website statistics tools also have different ways of referring to the traffic. Let's look at some common definitions, starting with 'hits'.

Simplistically, a hit is counted for each and every piece of information that is displayed on your website. (A more technical explanation would be that a hit is counted for each information request received by the server). So if you have 5 pictures on your page, there will be a hit counted for each of those pictures. You can see that hits are not a very accurate way of measuring website traffic.

A better measure would be Unique Visitors. A unique visitor is, in theory, one person visiting your website during the reporting period. Now, the Internet cannot 'see' you. It doesn't know that you are the same person using your PC from home and from work. The way that the Internet measures a 'unique visitor' is by looking at the IP address of the requestor (that is normally you). When you are logged into the Internet you get assigned an IP address. Each server on the Internet has an IP address. IP addresses (Internet Protocol Addresses) is like a street address, or a telephone number. Now, the problem is that there is really no consistent way of assigning IP addresses. In most of the rest of the world, an IP address is pretty much permanently allocated to you when you get an Internet connection. 

Not so in South Africa. Most of our traffic is routed through Telkom, and Telkom has big blocks of IP address ranges that it assigns to other service providers. This means that the chances of you ever having a unique IP address is very slim. You might share your IP address with various other Internet users. Also, the majority of website users in South Africa makes use of their work Internet facilities. Places of work are normally connected to the Internet through a proxy, with one IP address. In practical terms therefore, you might have 200 workers all using the Internet through one IP Address. 

The concept of Unique visitor therefore, also does not spell out the full picture.

You are also going to need information about Visits. One unique visitor might visit three pages. Or return to visit your site after an hour. Some website statistics programs also collect information about the Pages.

The funny thing is, not any one website statistics program will give you the same statistics. They all measure website statistics slightly differently. For example, one program might take the time period between visits to be one hour, others might only see it as a half hour. In other words, if someone visits a website and stay on a page for 20 minutes, then close a page and return to it within a 30 minute timeframe it will be counted as 1 visit. But if someone visits a website, close a page and return to it after 1 hour and 10 minutes, it is seen as 2 visits. See why it is not so straightforward as it sounds? My recommendation is to find a package that you are comfortable with and stick to that, so that you always have a constant base to base your frame of reference on.

Some popular web statistics packages

Most web hosting companies provide web statistics included in their hosting packages. At least, the better ones do. Watch out for any hosting company that wants to charge you extra for website statistics.

The two most popular web traffic statistics packages (at least on the Linux platform) are Awstats and Webalizer. Let's take a look at them both. (In a next installment I will go into which statistics tools you can use if your hosting provider does NOT provide you with either Awstats or Webalizer - look out for it)

Of the two, I prefer Awstats. This is the main layout of the Awstats stats. 

 

As you can see, they track Unique Visitors, Number of Visits and Number of Pages. They also make provision for breaking these stats down to a daily basis, as well as giving you geographic breakdowns, breakdowns by browser type, breakdowns by referrer (in other words, from which other websites did people click through to your website). It also has a full list of keywords that were used to find your site.

This is the layout of Webalizer:

As you can see, although these statistic were drawn from the same website, on the same date, the statistics look quite different! While we also have visits and pages, they do not correspond to the Visits and Pages we have in Awstats - possibly because they measure them differently, as previously discussed. While Webalizer also give you breakdowns by country, by browser etc, they only give you the top 25 or so keywords used to find your site. To me this is a drawback since you would want to study all your keywords to see whether your marketing efforts are paying off, that people are finding your website throught the right keywords etc.

The more visitors that come to your website, the more accurate your interpretation will become. The greater the traffic is to your website, the more precise your analysis will be of overall trends in visitor behavior. The smaller the number of visitors, the more a few anomalous visitors can distort the analysis.

What can you learn from your website statistics?

The aim is to use the web traffic statistics to figure out how well or how poorly your site is working for your visitors. One way to determine this is to find out how long on average your visitors spend on your site. If the time spent is relatively brief, it usually indicates an underlying problem. Then the challenge is to figure out what that problem is.

It could be that your keywords are directing the wrong type of visitors to your website, or that your graphics are confusing or intimidating, causing the visitor to exit rapidly. Use the knowledge of how much time visitors are spending on your site to pinpoint specific problems, and after you fix those problems, continue to use time spent as a gauge of how effective your fix has been.

Additionally, web traffic stats can help you determine effective and ineffective areas of your website. If you have a page that you believe is important, but visitors are exiting it rapidly, that page needs attention. You could, for example, consider improving the link to this page by making the link more noticeable and enticing, or you could improve the look of the page or the ease that your visitors can access the necessary information on that page.

If, on the other hand, you notice that visitors are spending a lot of time on pages that you think are less important, you might consider moving some of your sales copy and marketing focus to that particular page.

As you can see, these statistics will reveal vital information about the effectiveness of individual pages, and visitor habits and motivation. This is essential information to any successful Internet marketing campaign.

Your website undoubtedly has exit pages, such as a final order or contact form. This is a page you can expect your visitor to exit rapidly. However, not every visitor to your site is going to find exactly what he or she is looking for, so statistics may show you a number of different exit pages. This is normal unless you notice a exit trend on a particular page that is not intended as an exit page. In the case that a significant percentage of visitors are exiting your website on a page not designed for that purpose, you must closely examine that particular page to discern what the problem is. Once you pinpoint potential weaknesses on that page, minor modifications in content or graphic may have a significant impact on the keeping visitors moving through your site instead of exiting at the wrong page.

TrackBack URI for this entry

Comments (0)