Probably the most basic of questions, but I'm just starting to learn this "web" stoof!

So, how do you track how many times a certain file has been successfully downloaded from your server? I have a stats page that comes with the web package, but it only gives broader stats.

Chatting with the tech support, they told me I'd have to install a script or something.

Can anyone clue me in on this?

I don't know a lot about the technical nature of these things, but this is tracked in my web server logs. Basic traffic stats tools don't usually show this, I think. I can get the data you are asking about if I download my web logs and analyze them with Download Analyzer (http://www.downloadanalyzer.com/).

-Tom

There are several ways to do this and they all have varying accuracies:

1. Count the number of hits on your file using a standard weblog analyzer. Very inaccurate because of download managers, broken downloads etc.

2. Take the total traffic to your site (from a standard weblog analyzer), divide that by the size of your demo and subtract a percentage for the web pages and broken downloads. Ie. Demo size is 1MB, total traffic is 1GB (1000MB) = 1000 downloads. Subtract 15% = 850 downloads. This is a good rough measure because the traffic created by downloads will be large enough compared to the basic site traffic to make this pretty accurate. The fuzz factor (15% in this case) is where the inaccuray lies, but you still get much better estimates than the first case.

3. If you have a way to track unique users you might want to track the number of users who initiated the download. In this version the download file is actually a script that logs the download attempt into a database and then outputs the file. This version requires you to do server side scription, both to be able to track unique users and then to link them to downloads. This version doesn't take into account broken downloads, but you can still subtract them using a standard factor. It is more accurate then 2 because you are able to factor out the hits on the web pages, and it enables you to track how many of your visitors download (visitor to download conversion).

In my opinion, number 1 is pretty useless. I personally use 3 as basis for the rest of my measurements. In the short run it is probably more important that you pick a single adequate measure and then keep using that as your basis for optimization. What you are interested in is deviations to things like conversion rate. You just get some initial measure and then observe how your changes are affecting that measure.