I host a lot of software products on my server (.exe files). The vast majority are downloaded via a direct link from a 3rd party site. For example a freeware site lists one of our screensavers on their site and the download link is a direct link to the file on our server (http://www.oursite.com/downloads/ourfile.exe). Once the user installs our software, we load a "Thank You" page. We use this to track the number of installs that each of our software products gets. However, when I look at our Webalizer stats for these .exe files, I'm confused. The example below is typical of the Webalizer stats for all of our .exe files. Hits: 8021 KBytes: 8050776 The filesize of the .exe in this example is 4445 KB. If I divide 4445 into 8050776, I get 1,811. So, in theory, if EVERY initiated download is fully completed, I should have had this .exe file downloaded 1,811 times. Obviously that is not realistic as not all initiated file downloads will be completed. The Webalizer "Hits" number of 8021 is evidence of that. However, I think you should be able to get a rough idea of the completed download number by dividing the filesize of the file into the total KB of bandwidth indicated by Webalizer (8050776 in this case). That number is 1,811. Realistically, due to incomplete downloads, the actual number of completed downloads might be, what, half? Let's say 900. The problem is, my count of installs shows closer to 200. I understand that number won't include downloads that don't actually get installed or instances where the post install "Thank You" page doesn't get registered. Still though, there seems to be a considerable difference between the amount of bandwidth I'm using (and paying for) and the actual resulting software installs that it generates. Any idea why there might be such a discrepency? Perhaps I'm just misunderstanding how Webalizer works and reports.