Home
Articles
Glossary
SiteMap
About
Email

Website activity statistics do not tell the whole story, introduction article.

You may have invested considerable funds on your company website and are likely to want to be able to monitor its performance.

There are a number of ways you can monitor website performance.

The more you invest in your website, the more you should monitor its performance.

To be sure you get results you expect and may have paid for, you should collect and analyse data which indicates performance and can highlight changes to make to improve or ensure your website's continued performance.

Graphical Statistics packages

Most commercial hosting packages include a statistics package in the price which allows authorised users to log in and see activity levels, in graphic format, on their website.

These graphs are usually built on the server from the raw access logs which servers maintain, recording all activity on the server and all files requested by users.

Many of them generate graphs or tables which offer considerable detail into: activity levels of users,

where and when they come
which pages they look at
which pages are the most popular
which are the most used entry pages
which are the most used exit pages
where do they come from,
which search engines or directories
which keywords they found your website with
how much data transfer your website has used
and much more.

Common examples of server located stats packages include:
Urchin
Webaliser
Analog
If you are unsure about what is available, contact your hosting company.

Counter

You can install a counter on your website pages, instead or as well as, using a graphical stats package. It is normal for commercial websites not to display a numeric count even if they have a counter installed as this is seen as somewhat amateur.

Even if you have a host based graphical stats package you may still decide to fit a counter because the website graphical statistics may not tell the whole story.

Normally counters are used if no graphic stats or logs are available for example where free hosting is being used.

Where counters have advantages.

Some search engines will cache your pages, Google (www.google.com) for example does this.

Google keeps a copy of your page in its memory and serves that copy to visitors if they request it.

People searching for suitable websites in Google will often look quickly at the page google has cached and may not actually visit your website or request any files from your server.

A JavaScript counter installed in the page will still count these views as page views while your website statistics may miss them.

Keyword Ranking and search engine position

Either manually or using a program you can check the position of your website on your target search engines for your target keywords.

Hopefully you will have spent some time before constructing your website and writing your copy deciding which words and phrases you are going to target to enable people to find you in the search engines.

Obviously your name will be one of them but other descriptive words are likely to be important.

The challenge involved in getting significant keyword ranking should not be underestimated as depending on the words you choose there may be considerable competition all of whom would like their websites displayed at the top of the first page in response to as many searches as possible.

Examples:

1. This website is called sticky marketing (the domain name being www.sticky-marketing.net)
on 20/10/2001 I took a quick look to see that links to this site are at positions 3 and 6 on Yahoo.com webpage matches and 3 & 7 on Google.com while it is at position 2 on northernlight and lycos.co.uk

Though those are not too bad it is the name of the website and people are unlikely to be searching for it directly so no real benefit in terms of visitors.

2. I recently added a glossary of terms which the search engines have started to list.

I checked on 20/10/01 for "marketing glossary"

Yahoo webpage matches list 154,000 pages referring to this on which mine appears at position 144, not bad but well past pages one and two where it might be found and visited so no use to anyone.

Google lists 389,000 pages referring to "marketing glossary" and my first page is listed at position 173 again not bad compared to the total but no use to me.

Interestingly excite.com lists 1,345,235 pages referring to "marketing glossary", the pages have not been submitted to excite so I am just looking at search term popularity here.

To do well with the term "marketing glossary" I will have to find a way to beat these other pages to the first two and ideally to the top of page 1.

3. Checking for the expanded term of "sales & marketing glossary" gives better results.

Yahoo lists 88,600 pages, up from 38,700 when I looked about two months ago, my page is at position 28 which is not too bad and may not take so much work to get onto the first two pages.

Google lists 161,000 pages, up from 133,000, and mine ranks at position 31, again room for improvement but not a complete disaster.

So I can see that trying to get a better ranking I could do better targeting the phrase "sales & marketing glossary" as there is less competition than for simply "marketing glossary".

What this does not tell me is how often these terms are searched for, more about this on another day.

Getting a change in your position for a keyword takes time, you have to research your target search engines, plan what specifically to change on your website, do the changes, ensure your website is re indexed by the relevant engine, check your position again and repeat until you get the result you seek.

Your ranking is affected by others so getting to the front page is a start but no guarantee you will stay there.

To give examples of the importance of getting good position on your key terms.

A site I am responsible for achieved good search term position in a key engine, driving 500 visitors from that engine in a single month. Those visitors went on to look around the website and viewed many more pages. In that single month page views went from an average of below 30 a day to an average 150 pages viewed each day.

Another website rose from nothing to 40 individual visitors per day (visitors not page views) most arriving from search engines rather than inward links which are also important.

Raw Logs

Graphical statistics do not tell the whole story.

Raw logs are the data recording activity on your website server.

They record the transactions visitors make with your server, which files they request, where they request them from and the such like.

Extended logs contain a great deal of data and can become quite large but interrogating them will give you plenty of additional information than is typically displayed by the standard outputs from your server graphical statistics program.

Examples include the amount of data transfer on each request which allows you to determine if the user received all the file they requested or not.

I recently found an example of a company who thought from their graphical statistics that a particular part of their website was very effective.

On looking into the raw logs I found that in some instances users had tried 9 times to load a page and on each of these tries they had failed to receive the page and had pressed the refresh or stop and reload buttons on their browsers.

The simple graphics stats displayed all these 9 clicks as accesses of these pages.

They were in fact 9 fails rather than successes and the section was in fact not performing nearly as well as the company thought.

Search engine indexing activity

Raw logs can tell you when the search engines actually index your website.

They often request pages faster than a counter can pick up.

Their page requests will appear in your graphical logs and you may picky up their URL or identity but using raw logs you can definitely tell:

which search engine visited your website
which pages it requested
how often they come
if they obey your robot commands

Enquiries

A primary reason for having a business website, how many enquiries are you getting? How many of these are resulting in business or are they not commercial?

Perhaps the key reason you want to have a promotional website is to generate business, can you tell if your website is generating business?

Author Mark Abraham (mark@sticky-marketing.net) 19th October 2001