• http://www.greenlaneseo.com billsebald

    > Are those two worlds about to collide?

    They haven’t? I thought they did. At least they did in my bubble. Good article!

  • Michael_B

    I would say the one area in which you will find compelling evidence to prove last click attribution is sub optimal would be with branded conversions. I run one site that over 50% of branded conversions are from non branded keywords, not knowing, or appreciating this would of held this campaign back from the volumes it has been able to achieve. Websites that drive sales that have a tendency to be highly latent also seem to benefit from moving beyond last click attribution.

  • Andrew Goodman


    That’s contrary to what Wister’s data seem to show.

    They don’t call them attribution wars for nothing :)

    And I emphasize that more of us need the tools to check our own multiple attribution influences, not just someone else’s.

    By disparaging the importance of last clicks and seeking to replace them with something potentially much more tenuous, I believe we’re wading into a lot of potential error. More next time.

  • http://www.cirrus-media.com andrea-wasik

    Great writing, interesting topic, love the whole nerd vs jock thing, sounds sterotypical but in my experience it’s been true. Nerds will win this war. Data and logic will defeat flash and fancy-talk.

    Can’t wait for the next installment.

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

    Andrew, great piece as usual.

    We’ve been saying this for a long time. Alan presented our first publicized research proving that the “buying cycle” was overblown at SES in Feb 2006.

    At SMX East in the PPC Buying Cycle panel I raised the question: even if it does happen, what exactly was accomplished by the first visit if the person didn’t buy, didn’t bookmark your site, didn’t search for you by name, but instead searched on another competitive search term? Unless someone can prove that that first visit creates a greater propensity for the person to click on your ad rather than someone else’s, I’m not sure the first visit accomplished anything.

    I do think it’s important to credit the competitive search phrase in cases when competitive search is followed by a brand search. It’s not terrifically common. The most we’ve ever seen was something like 8% of last touch brand actually being preceded by competitive search. But that can be material. We give credit preferentially to the last non-brand/competitive touch.

    Give ’em hell, Andrew!

  • http://searchmarketingcommunications.com Tim Cohn

    The anti-last click camp(s) are just shills for display ad sellers.

  • http://mattlillig.blogspot.com mlillig

    Yahoo! has had a multi-channel attribution report, within the Yahoo! Search Marketing platform, called ‘Assists’ for about 3 years now. Beyond Conversions and CPA, of course, the Assist is probably the most popular metric for our advertisers right now. It’s helped our advertisers a lot in trying to understand how their upper funnel keywords drive conversions for the lower funnel keywords. So rather than lowering the bid for specific keywords because they thought they were poor performers (via direct response, last-click metrics), they’re are now able to measure that those keywords are valuable at driving conversions for their other keywords. It’s also helped our display advertisers in measuring how clicks on their display help to drive conversions to their other campaigns (such as search)..

    Google, I know, has also released new attribution reporting:

    Better attribution reporting is leading to more spend on both search and display and it helps to prove the value of search + display convergence.

    And it doesn’t have to stop there. Attribution tools can benefit advertisers for measuring email + display + search convergenceor shopping ad + search convergence or email + display convergence, etc.

    Attribution tools are making advertisers smarter about how they manage their online ad budgets. Sure, you can continue to measure last-click conversions, but you’ll be basing your budgeting decisions on less actionable data.

  • http://www.fuelinteractive.com/ briancarter

    Brilliant, Andrew.

    I’ve been looking forward to better analytics to visualize the touchpoint cycle but hadn’t occurred to me- the same diversity of conversion funnels we see inside some sites will happen with multi-site funnels.

    Very good point.

    Which third party PPC co is it that allows you to use a slider to favor early vs. late click attribution? Can’t remember. Have you used it? LOL sorry for the vagueness :-)

  • DavidSH

    I think the problem is that we (as Digital Marketers) are confusing “influencing” with “conversion”.

    I noticed another writer on this site citing the 2009 Forrester/iProspect study that showed 27% of users who “responded” to a Display ad performed a Search. That is nice, but what she failed to mention was that only 14% of all searchers who first learned of the product/brand via Display actually converted via Search (or as the study puts it “eventually”), interesting but hardly compelling.

    Of course looking at multi-touch points is important but bottom line its about the sales, Search is a bonifide sales channel, users still need to be convinced that the client’s offering is what they need. This “bottom funnel” view of search is a bit myopic, especially when you throw non-brand into the mix.

    To me the more compelling question to ask is what channels influence others and what can we do as Marketers to compel people to have greater interest/desire so that they will convert higher whether within their own channel or in another. If someone needs to find more information after viewing a display/search ad (or TV for that matter) then that channel FAILED to do what it was supposed to do and it should not get all the “credit”…but it should be viewed as influential and a potential opportunity for campaign improvement.