Lies, Damn Lies, Statistics, and Website Hit Counts

No one really knows how many people are connected to the Internet these days but one good estimate is about 300 million. How many of these are trade unionists? Here one would only be guessing, but the number is certainly in the millions and probably tens of millions.
From what I have seen so far, trade union websites are generally not reaching union members online. They often reach astonishingly small numbers of people.

One of Britain’s largest unions, with more than a quarter of a million members (most of them connected to the net) has a website which is regularly visited by less than one percent of the membership, according to statistics I’ve just seen.
For union websites to be effective, they must reach large numbers of members of those unions — indeed, they should aim to reach every member who is connected to the net. In addition, they are also often targetted at potential members, members of other unions, journalists, politicians and the broader community. Websites for teachers unions might target parents and children as well.
It’s important to know if our websites are being used for a number of reasons. To get unions to invest in websites (including hiring professionals to serve as webmasters) one should be able to demonstrate that they are being accessed. Over time, unions will want to know if certain print publications (such as internal newsletters) can be replaced by online versions. To make informed decisions, unions need to know if members use the websites which they’ve created — often at great expense.
How do we measure the success of a trade union website? Very often we hear about “hit counts.” Webmasters will sometimes boast of very large numbers of hits — thousands per day, for example — as evidence that the site is in fact being used and read. To find out whether this is true, one has to understand a bit about how things are actually counted.
Every access to a website will usually generate a record on the hosting computer. These records are stored in a log file which can be accessed using various kinds of computer programs which interpret the raw data. Three numbers are often produced by these programs: hits, page views, and unique visitors.
Hits is almost always the largest number and it’s sometimes very large indeed. But hits is a raw count of every single file of every type downloaded from the web server. This means every little graphic image as well. A web page which is heavy with graphics can produce 10, 20 or more hits each time it is accessed. Hit counts are, therefore, usually wildly inflated numbers.
Page views are far better and usually show how many times web pages (what are called HTML files) are downloaded. Sites that have lots of content will have high page views even if the number of visitors is quite small.
The best number we have is the third one — unique visitors. (This is sometimes called hosts or IP addresses). This counts the number of individual computers which have accessed your website. Every computer connected to the Internet is assigned a unique address, so in theory this would count the actual number of visitors to the site.
However, computers which are not permanently connected to a network — for example, those with modems which dial up to obtain Internet access — are assigned new addresses each time they connect. (This is called dynamic IP addressing.) A trade union member accessing his union’s website every day using a modem may be counted as 365 unique visitors in one year.
On the other hand, dialling up through an internet service provider usually means that web pages are filtered through their proxy servers. What does this mean for your count? If 1,000 members of your union connected to your through such a proxy server, they might all be seeing the web page as it is stored by the internet service provider. You’d be getting a count of only one unique visitor (when there were really a thousand).
Hit counts are worthless, page views only slightly better, and unique visitors are distorted by dynamic addressing and proxy servers.
So how do you know how many of your members actually visit the website and use it?
Here are a few ways:
1. Get members to register on the site — perhaps to enter member-only sections. (But don’t make anyone register to use the public part of the site.)
2. Collect their email addresses using a form on the website.
3. Do a survey — offline, perhaps using the union’s print publications — asking them if they use the site and what they think about it.
4. Ask for a show of hands at the union’s conference.
Be creative. Just don’t rely on the “hit count” to find out if members are using your website.