Web statistic discrepancies

[roblasch]

I have a website that uses ad serving software to track traffic that is sent to our advertisers sites. Often there is a discrepancy in the amount of traffic recorded between our software and the software used by the advertiser. I am confident that our tracking software is accurate because I can test it in real time. You can click on an ad and see the click in a report. I am looking to find EVERY possible cause for a click to be missed by the software analyzing traffic on a website (whether it is javascript based or a log file analyzer). Any and all input is greatly appreciated.

 

[xutopia]

does your ad server count click-throughs or just clicks? The difference could mean a lot.

Gary

http://www.xutopia.com

Haran
==========================

 

[ChrisHunt]

Well, the first thing that occurs to to me as a reason for the mismatch is caching. If the advertiser's site is stored in a cache somewhere, it could be served to the customer without the request ever going to their server and being logged.

No doubt there are other possibilities...

-- Chris Hunt
Extra Connections Ltd

http://www.extracon.com

 

[roblasch]

The ad-server says it is counting click-throughs. What would be the difference? It is an ASP program, so it goes to an ASP page when someone clicks on an ad, then records the click and redirects to the advertisers page. I am aware of the caching, and that could certainly acount for some of the missing clicks. But the discrepancy is around 40% which seems high for just caching. The makeup of our site makes it seem unlikely that alot of users return to our site and click on the same ads over and over. So if anyone knows of other possibilities, please let me know. I am trying to compile a complete list of anything that could cause this discrepancy.