The Magic & Wizardry Of Attribution

In my house, we spend random conversations speaking in British accents, talking about and acting like we’re characters in “Harry Potter.”  I know -- this might sound strange to you, especially when you realize I’m not from London; I’m from upstate New York.  My boys like it, though. They can’t wait to be old enough to read the books and watch the movies!

What doesn’t seem to work as of yet is the magic and wizardry behind attribution.   Is there a chance that attribution will sometime soon become the realm of muggle science, or will it always be a fantastical goal?  Will it be our industry’s Moby Dick?

Attribution is a key challenge for marketers -- probably one of the most important ones.  Just as a solid measurement methodology was back in the late 90s and early 00s, attribution is a hurdle for the industry. If we can master it,  budgets will continue to shift and advertising will become an even-more defensible category for brand marketers. 

When I speak of attribution, I am speaking of two areas:

1.     Online attribution (historically, last click vs. attributed clicks)

2.     Online vs. offline attribution (the effect of online plus offline in unison)

The online attribution challenge is the first one we need to tackle, as it is the easier of the two, if you can believe it.  How do we attribute value to each messaging opportunity in the continuum of online ads that a user is exposed to -- and how do we ensure that this value is fair and acceptable?  

To date, there are many different ways agencies and brands have calculated attribution and there are just as many ways publishers and networks game the system.  They are cookie bombing, jacking, plugging and pummeling (any combination of aggressive, relentless words) their way to being valued for their exposure, and in the end marketers and the consumers lose.  

These unethical practices remind me of the days of pop-ups and pop-unders, when ads were served to count an impression, even though the chances they were seen were minimal at best.   The good guys lose because they get lost in the clutter of ads, and consumers get untargeted, frequency-busting ads delivered right to their desktop. 

In the end, some industry group is going to have to get involved and help deliver the parameters for attribution, if not the actual attribution model itself.  I don’t think they can go so far as to authorize a single model, but the parameters for counting and ad delivery could indeed be set as a standard.

In the end, I foresee a handful of attribution models becoming industry-accepted norms, considered table stakes for any player in the business.  Attribution will depend somewhat on the goal action and the desired balance between brand and direct-response objectives.  A direct-response advertiser may have an attribution model more heavily weighted toward the latter part of the exposure stream, while a brand advertiser will be more weighted evenly through the exposure stream.  The final decision will lie in the hands of  marketers, who will also be tasked with conveying that model to each of their partners.

Assuming we can figure out the online attribution issue, the online vs. offline model becomes true wizardry.  In the offline world, you lack the ability to tag and calculate actual exposure for your messaging, and this creates a deeper challenge.  I don’t know the timetable for anyone to develop an accurate model for cross-channel attribution, and doing so is far more a mathematical model than a provable parameters-driven solution.  If I were to guess, I would say online attribution will be finalized in the next two years but cross-channel attribution may take more like 10 years.  I hope that’s not the case, but we will see.

With that, I go back to my “Harry Potter” movies, gazing into the mirror in the Chamber of Secrets and wondering what things will be like when attribution is a standard piece of the business.  It’s a promising future if we can figure it out, don’t you agree?

Tags: attribution
Recommend (9) Print RSS
6 comments about "The Magic & Wizardry Of Attribution ".
  1. Jon Mandel from Precision Demand , November 28, 2012 at 11:11 a.m.
    Cory-It isn't wizardry. PrecisionDemand has been doing offline and online attribution for three years. And not through wizardry mumbo jumbo but real sales effect measured at the cash register/order page and real media...not theory. But Swiss precision, Swiss neutrality. It isn't wizardry it is hard science.
  2. Eric Wittlake from Babcock & Jenkins , November 28, 2012 at 12:03 p.m.
    Hi Cory, Yes, attribution is definitely a complex topic. One of the challenges, in my opinion, is that media companies (particularly networks) learned very quickly how to game last click and last impression attribution. As attribution models "improve" the method to game attribution will need to change, but what will keep media companies from continuing to game it? Because of this, I'm not convinced that your #1 will actually be the easier proposition to solve for large marketers. Large brands have, for years, used complex models to determine the impact of marketing investments and these are not multi-million dollar projects anymore as more organizations and individuals have experience with it. We will see what happens, however the measurement I want to see the industry move towards is lift (the change marketing drives), not attribution. When we start measuring the change marketing creates instead of the activity that can be matched back to marketing using cookies, the industry will be in a better position to win the larger traditional media budgets you reference. -- @wittlake
  3. Steve Climons from Crossover Marketing & Advertising , November 28, 2012 at 3:49 p.m.
    Cory, Isn't attribution really the process of how you measure the net results from your campaign? If the online campaign through clicks or in combination with offline campaign url push attributed to traffic and sales then the campaign could be deemed a success. I think the magic or wizardry is the campaign messaging and how effective it was in engaging people and not annoying them.
  4. Matthew Snyder from ADObjects Inc , November 28, 2012 at 5:56 p.m.
    Cory! I really like the piece as you are right. As long as channels are so disconnected, building models of Attribution takes quite a bit of coordination and strategy to set-up and get right. We have been working hard on this problem for quite some time as well and look at it from a new perspective from the ground up as to what needs to be done to create the right campaign that can solve this attribution problem. A new way to re-imagine advertising is fundamental...... Responsiveads.com
  5. Lindsay Durfee from RevTrax , November 29, 2012 at 10:24 a.m.
    Cory, I agree that attribution is a critical KPI for marketers. An interesting challenge our company is tackling is connecting the dots between online and offline - rather than analyzing online vs. offline in silos. For retailers and consumer product companies running multi-channel online promotions (email, social, digital, etc.), how many of them can point to a revenue trail that ultimately leads to a specific online consumer showing up in-store and making a purchase? Providing this data for marketers allows brand teams and agencies to better allocate budget, improve creative/media buys and, most importantly, drive sales. If you're interested in the details, head over to www.revtrax.com to see how we're doing it. Thanks for starting the discussion.
  6. Kenneth Hittel from Ken Hittel , December 1, 2012 at 12:56 p.m.
    Rarely has so much digital ink been spilled over such a fruitless subject matter. Attribution is a bunk problem that will never be "solved" -- not in the next two years, not in the next ten. In the real world, buying decisions are always (and necessarily) over-determined, in much the same respect as Freud used the term when discussing what might be called the attribution of mental illness. Looking for the one deciding fact or action that leads to a buy, or trying to unravel every step on the long and winding road that culminates in k-ching, is a total waste of time. The task is an impossible one. Let's just be content with what we can point to: that click on the PPC ad, the 800-number or "Buy Now" button on the Web site, the tagged ad campaign delivering people directly to the sell, etc. For all the rest let's use metrics appropriate native to the activity, un-borrowed from traditional & limited ROI.