Analog 6.0: How the web works

This section is fairly long, but it's worth reading carefully. If you understand the basics of how the web works, you will understand what your web statistics are really telling you.

So, what do you know about it? First, I make one request for your front page. You know the date and time of the request and which page I asked for (of course), and the internet address of my computer (my host). I also usually tell you which page referred me to your site, and the make and model of my browser. I do not tell you my username or my email address.

Next, I look at the page (or rather my browser does) to see if it's got any graphics on it. If so, and if I've got image loading turned on in my browser, I make a separate connection to retrieve each of these graphics. I never log into your site: I just make a sequence of requests, one for each new file I want to download. The referring page for each of these graphics is your front page. Maybe there are 10 graphics on your front page. Then so far I've made 11 requests to your server.

After that, I go and visit some of your other pages, making a new request for each page and graphic that I want. Finally, I follow a link out of your site. You never know about that at all. I just connect to the next site without telling you.

The other sort of cache is on a larger scale. Almost all ISPs now have their own cache. This means that if I try to look at one of your pages and anyone else from the same ISP has looked at that page recently, the cache will have saved it, and will give it out to me without ever telling you about it. (This applies whatever my browser settings.) So hundreds of people could read your pages, even though you'd only sent it out once.

You can also know what people told you their browsers were, and what the referring pages were. You should be aware, though, that many browsers lie deliberately about what sort of browser they are, or even let users configure the browser name. And some people use "anonymizers" which deliberately send false browsers and referrers.

1. You can't tell the identity of your readers. Unless you explicitly require users to provide a password, you don't know who connected or what their email addresses are.
2. You can't tell how many visitors you've had. You can guess by looking at the number of distinct hosts that have requested things from you. Indeed this is what many programs mean when they report "visitors". But this is not always a good estimate for three reasons. First, if users get your pages from a local cache server, you will never know about it. Secondly, sometimes many users appear to connect from the same host: either users from the same company or ISP, or users using the same cache server. Finally, sometimes one user appears to connect from many different hosts. AOL now allocates users a different hostname for every request. So if your home page has 10 graphics on, and an AOL user visits it, most programs will count that as 11 different visitors!
3. You can't tell how many visits you've had. Many programs, under pressure from advertisers' organisations, define a "visit" (or "session") as a sequence of requests from the same host until there is a half-hour gap. This is an unsound method for several reasons. First, it assumes that each host corresponds to a separate person and vice versa. This is simply not true in the real world, as discussed in the last paragraph. Secondly, it assumes that there is never a half-hour gap in a genuine visit. This is also untrue. I quite often follow a link out of a site, then step back in my browser and continue with the first site from where I left off. Should it really matter whether I do this 29 or 31 minutes later? Finally, to make the computation tractable, such programs also need to assume that your logfile is in chronological order: it isn't always, and analog will produce the same results however you jumble the lines up.
4. Cookies don't solve these problems. Some sites try to count their visitors by using cookies. This reduces the errors. But it can't solve the problem unless you refuse to let people read your pages who can't or won't take a cookie. And you still have to assume that your visitors will use the same cookie for their next request.
5. You can't follow a person's path through your site. Even if you assume that each person corresponds one-to-one to a host, you don't know their path through your site. It's very common for people to go back to pages they've downloaded before. You never know about these subsequent visits to that page, because their browser has cached them. So you can't track their path through your site accurately.
6. You often can't tell where they entered your site, or where they found out about you from. If they are using a cache server, they will often be able to retrieve your home page from their cache, but not all of the subsequent pages they want to read. Then the first page you know about them requesting will be one in the middle of their true visit.
7. You can't tell how they left your site, or where they went next. They never tell you about their connection to another site, so there's no way for you to know about it.
8. You can't tell how long people spent reading each page. Once again, you can't tell which pages they are reading between successive requests for pages. They might be reading some pages they downloaded earlier. They might have followed a link out of your site, and then come back later. They might have interrupted their reading for a quick game of Minesweeper. You just don't know.
9. You can't tell how long people spent on your site. Apart from the problems in the previous point, there is one other complete show-stopper. Programs which report the time on the site count the time between the first and the last request. But they don't count the time spent on the final page, and this is often the majority of the whole visit.

Defenders of counting visits etc. claim that these are just small approximations. I disagree. For example, almost everyone is now accessing the web through a cache. If the proportion of requests retrieved from the cache is 50% (a not unrealistic figure) then half of the users' requests aren't being seen by the servers.

Other defenders of these methods claim that they're still useful because they measure something which you can use to compare sites. But this assumes that the approximations involved are comparable for different sites, and there's no reason to suppose that this is true. Pirolli & Pitkow's results show that the figures you get depend very much on how you count them, as well as on your server configuration. And even once you've agreed on methodology, different users on different sites have different patterns of behaviour, which affect the approximations in different ways: for example, Pirolli & Pitkow found different characteristics of weekday and weekend users at their site.

Still other people say that at least the trend over time of these numbers tells you something. But even that may not be true, because you may not be comparing like with like. Consider what would happen if a large ISP decided to change its proxy server configuration. It could substantially change your apparent number of visits, even if there was no actual change in the traffic levels at your site.

I've presented a somewhat negative view here, emphasising what you can't find out. Web statistics are still informative: it's just important not to slip from "this page has received 30,000 requests" to "30,000 people have read this page." In some sense these problems are not really new to the web -- they are present just as much in print media too. For example, you only know how many magazines you've sold, not how many people have read them. In print media we have learnt to live with these issues, using the data which are available, and it would be better if we did on the web too, rather than making up spurious numbers.

Another, extremely well-written document on these ideas is Measuring Web Site Usage: Log File Analysis by Susan Haigh and Janette Megarity. Being on a Canadian government site, it's available in both English and French. Or for an even more negative point of view, you could read Why Web Usage Statistics are (Worse Than) Meaningless by Jeff Goldberg.

Need help with analog? Use the analog-help mailing list.

[ Top | Up | Prev | Next | Map | Index ]