Msg#: 3659016 posted 7:55 am on May 26, 2008 (gmt 0)
Group, can you suggest me on how to track all the assist keywords used by ad-words visitor before actually filling the form. I mean the visitor may have visited many ad-words landing pages before finally filled form using a key-phrase. When we calculate the performance of a keyword, we tend to miss on the keywords that did not convert into a lead but were influential in the process. We want to track these in-house and not depend on analytics for the same..
Msg#: 3659016 posted 1:52 pm on Jun 3, 2008 (gmt 0)
To what I understood, "Assist Keywords" are the ones that do not actually have converting visitors on to the website. However browzer history of converting visitors reflects that these people have earlier visited the website using some other keywords. Now, these other keywords are "Assist Keywords". I hope, I am able to communicate this..
Msg#: 3659016 posted 2:48 pm on Jun 3, 2008 (gmt 0)
Okay, that's very clear. I don't have anything to recommend other than capturing all KW with their dates in a cookie, doing a dump at the time of purchase, and using some kind of regression program or neural net analysis to come up with the contribution of each. Or something much simpler and cruder can be used, like weighting based on time lag and # of intermediate KW.
One of the problems with this analysis is going to be comparing the KW patterns for purchases to the KW patterns of those who didn't purchase. If you can't do that, you can't get a handle on causality at all.
I know that big companies do this all the time and most of them use an ordinary stats program like SAS or SPSS. The data collection shouldn't be too hard. Getting an accurate modeling of the data will be.