Attribution Management Revisited

The piece on viewthrough and how to more accurately measure it in last week's Metrics Insider was appropriate from a timing standpoint. And as Paul Harvey was fond of saying on the radio: "Now, for the rest of the story."

Last fall, I wrote about how last click attribution is about to become a thing of the past. This has been called Attribution Management or Multiple Attribution Protocol. Both DoubleClick and Atlas have active testing going on in this arena.

So this weekend I read with great interest in BusinessWeek of Microsoft's plans "post Yahoo bid." According to Keith Lorizio. who heads up online sales for Microsoft, the company "is also making the case that search advertising, Google's gold mine, is overrated. In the months ahead, it plans to introduce new ad technology that it says will demonstrate that to advertisers. 'We're going to win with this strategy,' Lorizio says."

Seems to me that out of the obvious tools that Microsoft has at its beck and call, the E-Map Beta (Engagement Mapping) is probably that technology. Recent presentations by Atlas have talked about the "Last Ad Standard" wherein the last click or ad viewed gets 100% of the attribution for sale, download or whatever action the consumer took. In these presentations, they go on to discuss the obvious -- that consumers are reached multiple times, across multiple channels, over extended periods of time. This is what in the media world we know as continuity combined with frequency. We've always known that it takes a certain frequency level to ensure communication (generally 3-10x is the ideal in most media types, but your mileage can vary).

I have also heard both DoubleClick and Atlas say in numerous presentations that there is an over-20% increase in conversion among sponsored search clickers who were also exposed to display ads.

The effect of this, when properly modeled, is that the efficacy of search (and resulting efficiency) decreases by 20-25%, and Web and other vehicles increase by the same amount, when evaluating on a CPW basis (cost-per-whatever a client is measuring).

If initial beta testing holds for extended tests, will the result be that the monetization that advertisers are willing to give Google credit for is 20-25% lower? Will that impact Google? Or will the rise in number of searches that happens when a company is doing graphical advertising offset this? (See why Google wants to get into graphical advertising in a much bigger way?)

The Engagement Mapping or Attribution Management application could be a big win for Microsoft. And it could be a big win for all display advertising. It remains to be seen whether "all boats will rise" or whether this will be at the expense of search. As Mr. Harvey also used to say, "Stay tuned for page 2."
Tags: metrics
Recommend (2)
2 comments about "Attribution Management Revisited".
  1. Matt Lillig from Yahoo! , January 26, 2009 at 8:37 p.m.

    Don't count out Yahoo on the attribution efforts. :)

    Here at Yahoo! we simply give those contributing ads/campaigns what we call an Assist. We then give the last ad/campaign clicked before a conversion, a Conversion.

    So for example, if you're using Yahoo's Full Analytics solution (found free in Yahoo advertisers' search accounts) and you're tracking display ads and search ads for example, you will see Assists attributed to ads or keywords that contribute to the conversion of your other ads or keywords.

    So for example, if a visitor originally clicks on a display (that's being tracked with Yahoo's Full Analytics) and then later converts off a Yahoo search ad (also being tracked by Yahoo Full Analytics), then the display ad will receive an Assist and the search campaign will receive the Conversions.

    Now, as an advertiser, I know that my display ads are helping to drive conversions to my other ads/campaigns.

    Assists are extremely helpful when it comes to budgeting for display campaigns or bidding ons earch keywords. Here's why....

    Most people only look at the direct response of the ad or campaign when making budgeting decisions (basically just looking at conversions). No attribution view is being factored in. But with Assists, the attribution view is opened. For example...

    Normally I made have dumped a display campaign because it didn't convert well on its own. But once I factor in Assists, I may see that while the display ad was not converting well on it's own, it was very valuable at driving conversions for my other ads (such as search). This happens a lot!

    Same with search. I may lower the bid on a keword because it didn't convert well on its own. Using Assists though, I find out that the keyword was driving conversion for my other keywords. Now rather than loweing the bid or pausing the keyword, I continue to spend or raide the bid.

    Simply put, Assists change the way advertisers manage their online ad/campaign budgets.

    For more info, check out:

    http://mattlillig.blogspot.com/2008/12/yahoo-assistsput-money-back-into-your.html

    and

    http://www.ysmblog.com/blog/2008/11/03/measuring-search-and-display-for-success/

    Regards,

    Matt Lillig
    Yahoo Search Analytics - Team Lead

  2. Mark Hughes from C3 Metrics , July 23, 2009 at 9:20 p.m.

    White Paper with actual results of deploying Revenue Attribution available for download on C3 Metrics' home page here - http://c3Metrics.com