Requests vs. Downloads

[board]

We use Urchin to track our Web statistics. Recently, we have noticed a difference in the tracking of a pdf file. The file has been "requested" 20,000 times but "Downloaded" just 10,000. Does anyone know what the reasons for such a huge difference might be, and which number is more reliable? I would guess that the requested number doesn't reflect whether the user was actually able to open the file, but the difference is larger than I would expect.

[monkeythumpa]

It probably wasn't changed since the last request or you are blocking someone from downloading it. If you are trying to track it you might append the link with a random variable so that the server sends the content every time.

[topsites]

Could be because half of all Internet users connect via a dial-up connection and most of these folk refuse to wait the length of time it would take to download this type of file. Just a thought, but the actual data is not far off.
Peace out,
Pascal

[edited by: tedster at 7:33 pm (utc) on Mar. 14, 2005]
[edit reason] remove extended signature [/edit]

[TonyHodge]

There are usually multiple requests in a log file for a single PDF. When you look at the log file entries for PDFs, you may notice that usually the first request has a 200 status code but all subsequent request for that same PDF have a 206 status code. This is a continuation or partial request. These should all be counted together as one.

Make sure you use a tool that does this or your numbers will be wrong.

[teaperson]

In a rare bit of useful documentation from Urchin, there's a FAQ on this issue: [help.urchin.com...]

Basically, as someone else mentioned, you can have multiple page requests for the same multi-page PDF, but the download count will only be incremented once.

[board]

Just want to say thanks to you all for your very helpful responses!