Hit Counters

How do you measure success

How Popular are You?

On this scale the best result is to have a stream of Email messages from your response message option, praising your design and content. The ultimate is to receive an Internet Best Website award, normally only received by site authors who go for elaborate designs, sexy graphics and Java animation, and ignore the principles of good navigation and useful content. (Or am I being over-cynical, not having won one myself?)

Associated with this, many people are neurotic about their ranking in the Search Engine lists. Where do you rate this week. How many weeks have you been top. Why have I been pushed down the list by him. You can spend a lot of time worrying about this.

A cruder but more direct method of measurement is to have a Hit Counter.

Health Warning: Many people regard a hit counter displayed on a page as being a bit naff. Generally we recommend that hit counters should not be displayed but discreetly available to you if you know where to look.

A hit counter is a device for showing you how few people have visited your site, to quote the cynics! So much is this true, that there was even a website dedicated to saying so run by Mr Jenner at Wisconsin but it seems to have gone away; perhaps he was discouraged by the low number of hits he counted....

One of the common devices is to display a constant figure. Some people may believe you. Some people may care ...

How many hits will you get

1. If a visitor comes back after a gap of say 15 minutes, that is a new visitor.

2. If you do nothing very much about publicity, and you have a website with family photos you could get 7 visitors/month, and that is chiefly from you.

2. If your site is static and publicised via search engine you get 70/Month

3. If you change the site often and have some links to your site, and you are getting a few hits via the search engines, then suddenly it goes up to 700/month mostly from the search engine robots. This is what I get on the Goring-by-Sea website.

4. Try a bit harder then suddenly it is 7000/month and you look at the logs and find that most of them are ISP cache visits. I get this one the Waller website

5. To do better then you have to have a product which is wanted by a very wide audience. Getting them actually to buy is something else.

So how many visits are you getting. Don't tell me cos I don't believe you. The real test is how many significant Emails and response form replies are you getting.

Needs some CGI coding

There is a demo count on my homepage and at the top of this page.

The on-page counter is driven by SSI (server side includes) and depending on the ISP, this may only operate if the page name has a .shtml extension.

If there is Java or JavaScript on the page with the counter, it may happen that the counter is incremented twice, once on the first loading, and again when the JavaScript cuts in. You may not think this matters.

Third Party Hit Counters

And unless the off-page hit-counter is right at the top of the the page, visitors may click a link to another page, without knowing or caring whether the counter has had time to be triggered. We recommend that the counter is the last thing on the page, then if there is any delay in the counter mechanism, the visitors can still enjoy their visit to you.

This is a demo counter which would normally be at the bottom left on your homepage. This counter comes from http://www.ipstat.com and is probably as good as any; you can click on the image and get the stats for this page from IPStat. It shows number of hits, hits per day etc, browser and system analysis. It shows the website of the last few hits, you may find this interesting, but most of the hits come the larger ISPs like AOL who have a small number of IP numbers which they allocate to surfers on a lending library basis; and then they are allocated to someone else.

If you want a second recommendation here http://www.thecounter.com or http://www.fxweb.com are said to be good. For more details on hit counter providers go to http://www.counterguide.com/

Lies, Damn Lies and Hit Counters

And some surfers will jump into the middle of the website; I tell people about individual pages and they go straight there and avoid the homepage, and are thus not counted. Other surfers make repeat visits and do not pull down the first page because it is already held in their cache. On my site, because of these things the first page hit counter is recording only 25% of the visitors.

Worse. Some of the bigger ISPs cache popular pages and the visitors get to see your website by getting the files from the ISP cache and are not therefore counted. On the other hand, each of the big ISPs send their cache robot to your site every night and check to see if anything has been updated. That could be a lot of hits which are not interesting to you. And the search engine robots may pay you a visit now and again, and each of these will cause a lot of hits which are useful but distort the figures.

Note that an on-page hit counter will only record hits on that page. Some of the hit-counter services will provide these stats for any page on which you install a counter. This is probably over-engineered and not the way you want to go; not least because you have then to check each page to see what the page counter shows..

Server Statistics

The hits are also summarised so that they can tell me how many distinct hosts are served. That is, if someone comes in several times during the week, this figure counts those several hits as the one count. On some servers thay can tell you how many are repeat visitors and how many are new.

On this site the most popular page varies according to whather it is highlighted as new or interesting on the Homepage. Apart from this, in the past it has been as varied as, surprisingly, the background wallpaper page, and also the Guide to writing a CV for a job application. These are probably an accident of placing of the reference on a search engine rather than genuine interest; I hope the visitors coming in here stay to see other things.

An Alternative to all this nonsense