• http://www.calculatemarketing.com alanmitchell

    Hi George,

    Excellent analysis of using cookies to help measure advertising effectiveness. Thought I’d tough on a point you made in the section ‘Click to Conversion Curve’:

    “That suggests to us that the behavior of that group can no longer be explained by latency from exposure to the ad, and instead their behavior must be driven by something else. That’s the point at which credit should no longer be attributed to the ad.”

    I agree there must be a cut-off somewhere, but I think it raises two further points:

    1) Even if the visitor’s purchase was driven by something else, but PPC did the hard work in originally educating the visitor about the product/service, shouldn’t some of the value be attributed back to PPC? In other words, even though the PPC visit may not have led to a sale on its own, when combined with another driving force, it may have helped.

    2) Is it right to ignore the value of PPC just because it did not generate a sale? When I have limited conversion data, especially at a keyword level, I often like to look at which areas are getting people to view the ‘contact’ page, or view the ‘store locator’ page, or revisit the site again on a later occasion ( http://www.calculatemarketing.com/blog/techniques/intelligent-analytics-for-intelligent-adwords-management/ ). My theory being that if keywords were only optimised based on sales, then only brand and product-related keywords further along in the buying cycle would be seen as profitable, with other keywords which generate engagement and returning visits ignored.

    Cheers,
    Alan

  • http://www.rimmkaufman.com George Michie

    Alan, thanks for the comments! The branding value of a visit is certainly not zero, and every business should think carefully about how to value non-converting visits. It’s important to watch the multi-ad interactions to see the extent of the funnel. One shouldn’t assume that general term clicks lead to more specific term clicks — that’s a fairly rare pattern in the data in most cases, but measurement is the key.