Page caching affecting logs

Is this an explanation?

[fom2001uk]

One of our sites has a frequently updated homepage. Every week, they change a couple of the links (news stories) and we tag them to identify any click-thus.

According to our logs, we have an overlap of hits to these links across different months. There is no way that these same links are on the site on different months.

So, can it be that when some people visit the homepage, they get a cached version, so the old links still appear for them to click on?

I can't think of any other explanation. It's driving me nuts!

[fom2001uk]

Nobody in today?

[onlineleben]

Could be related to your click-tracking software.
If the links are setup like yourdomain.com/click.cgi?ID=targeturl it could be that this URL and its contents are indexed by the SEs and therefore remain visible longer than you want to.
Why worry as long as you get decent traffic that is also on target?

[aspdaddy]

Fom,

It could be returning visitors, getting cached pages from thier broswer not the proxy cache.

You could no-cache the pages but use extra bandwith doing so.

[lorax]

I believe it is a caching issue but the cache could be on any piece of equipment between your server and the client. You could use the NO CACHE directive but even that is not a guarentee.

[cornwall]

The major ISPs are not likely to cache for long periods. As I understand it, say AOL or Freeserve, will serve up a cached version to their subscribers but that will only be there for 24 hours before being refreshed.

Perhaps someone with more exact knowledge, can comment on cache time

[aspdaddy]

>The major ISPs are not likely to cache for long periods

This is why i think its the browser cache, if IE is set not to get the latest version, it will stay in there until you clear the cache.

[jdMorgan]

You can control this by setting up cache-control headers for each of your resources, or by resource type, i.e. .html, .gif, etc.

For Apache, see mod_headers [httpd.apache.org] and mod_expires [httpd.apache.org].

Example:

# Set up Cache Control headers
# Default - Set http header to expire everything 1 week from last access, set must-revalidate
ExpiresActive On
ExpiresDefault A604800
Header append Cache-Control: "must-revalidate"
<Files index.htm>
ExpiresDefault A3600
</Files>
<Files index.html>
ExpiresDefault A3600
</Files>
<Files news.html>
ExpiresDefault A60
Header unset Cache-Control:
Header append Cache-Control: "no-cache, must-revalidate"
</Files>

Also see:
Cacheability tutorial [mnot.net]
Cacheability tester [ircache.net]

HTH,
Jim