I understand that discrepancies will occur, but this is almost exactly double. What's more, these clicks are averaging about $2 each. There have only been a few conversions, and I added the Google tracking tool to the thank you page for those form submissions. However in each case, when we received 1 email from the form, Google showed 2 conversions. These things stand out cause there's not much traffic, and the 3 conversions have each happened on separate days.
I emailed Google and 24+ hours later got their boilerplate reponse of how we agreed to go by their metrics, etc., there can be reasons for the difference, etc. (The only filtered IP on WTL is me, and I haven't clicked on any of the ads.) It's very frustrating not to be able to talk to someone at Google (or even have a meaningful email exhange) about this subject.
So I'm posting this question here. I don't know what to do to evaluate the problem. Does anyone have advice? Is there a way to find out if this is double counting? Has anyone used WTL (or another JS tracker) and experienced the same issue? Should I just assume Google is right?
IP selection: Your tracking system may filter out visits from the same IP address; however, Adwords does not. If your user used a shared ISP provider (such as AOL or Earthlink) they may be sharing one or more IP addresses with other users. Therefore multiple clicks from the same IP address could be legitimate clicks from multiple users, and each one would be reported as an individual with your Adwords account but your tacking system might be unable to distinguish the users.
This means that users on dialup which tend to have 100's of users coming from one single IP lots of third party tracking programs will only see this as 1 person. These programs also overlap users so you only get the correct stats for the last user that entered your site using that IP.
I know it’s hard and frustrating to be putting your money into your new project and things not working out smoothly, but feel confident that Adwords is a tried and tested system.
