The Trouble With Legacy Attribution In An RTB World

How do you measure campaign success? If you’re like most marketers, you count clicks. And you engage multiple display partners to help drive those clicks, rewarding those who generate the last one before conversion.

The trouble is, the last-click attribution model works well (sort of) for exactly one type of campaign: driving sales among consumers who’ve visited your site. If that’s the only type of campaign you’ll ever launch, then your challenges are pretty manageable. But if your goal is to raise brand awareness or attract new buyers to your brand, then the attribution model will be sorely disappointing.

Most attribution approaches are stuck in the past, based on metrics and goals that are legacy to the brand, an issue all ad-tech providers face when working with large corporations. So how should marketers attribute credit?

Attribution should focus on individual campaign goals. When I talk to clients about attribution, I begin with marketing goals. What does each campaign need to achieve? Brand awareness? New customer acquisition? Reactivation? Once you define the goal, then you can decide where and how to look for success



I am an advocate for replacing legacy attribution models with ones that understand and reward all factors that contribute to campaign success, such as position-based attribution. This isn’t futuristic stuff; brands are successfully using position-based attribution today by measuring view-through, a metric that tracks the role of ad sequencing and frequency.

How View-Through Works

Every ad-tech partner, including programmatic platforms and ad networks, will pixel a consumer in three places. The first is the impression in which an ad appears, so the marketer knows who has seen it. The second is the ad’s click-though link, so marketers can track who engaged with it. The final is through strategic pixeling of the marketer’s website to measure consumer purchase intent.  Site pixels can link the media campaign to qualified interest in the marketer’s product.  A pixel on the conversion page, which is the page that says “Thank you for your order,” can ensure marketers are measuring the entire consumer path from initial display ad, to click, to site exploration to purchase.

The view-through allows marketers to connect those things together so they can understand how many ads consumers see before converting, and to attribute a marketing value to each RTB display ad, video ad or mobile ad that play a part.

Given all the talk we hear about the customer journey and driving better campaign return on ad spend (RoAS), why aren’t more marketers developing goal-focused attribution models? The truth is, people are reluctant to stray from their familiar models, and it’s up to the ad-tech vendors to educate their clients on better ways to value and attribute credit across all of their media partners.

At the moment, it’s pretty difficult to understand why, as a marketer, you need to change the way you measure what you do. And it’s even harder to explain it to your CFO who, as the ultimate numbers guy (or gal), is familiar and rather attached to last-click attribution.

And yet, we all know that in order to deliver on the real promise of RTB advertising, marketers must anticipate consumers. We need to identify and message our next customers while they’re in the pre-consideration phase -- because ultimately, that’s how to raise awareness and grow a business.

6 comments about "The Trouble With Legacy Attribution In An RTB World ".
Check to receive email when comments are posted.
  1. Chris Innes from SteelHouse, February 13, 2014 at 1:26 p.m.

    This is an interesting problem that is born from the behavior of paid search. As a consumer, I pull up a search engine when I'm looking for something and I act on those searches with a click.

    Display advertising is much different than the paid search model. Users are influenced with behaviorally driven display campaigns that shouldn't be judged by the click, first or last. The only true test of performance is a PSA control group with the goal of finding the incremental value of both click and view conversions.

    What's more interesting is that the TV world has already conquered this challenge - but translating that to a direct response world is difficult.

    I hope we can find ourselves in a future where more data is shared from each vendor / partner. Details such as the time between the impression and visit to the site, more multi-screen data, and true control groups can lead us to a place of better attribution that includes the view-through. Educating the marketer / CFO is the most difficult part of this task.

  2. Benjamin Masse from Triton Digital, February 13, 2014 at 3:03 p.m.

    Totally agree! We need to go beyond the CTR/last-click attribution. This KPI has limited usefulness especially from a branding perspective.

    Digital audio can also be tracked in a similar way with listen-through attribution. We've posted an explanation of this process on

  3. David Scacco from david scacco, February 13, 2014 at 4:32 p.m.

    This is good stuff. Marketers that find a better way to measure attribution will have a distinct advantage. Here's another way to approach this

  4. Susie Stulz from susie stulz consulting, February 14, 2014 at 4:34 p.m.

    I agree with Chris Innes that a lot of attribution is a carry over from search, which is really different from display. With search, folks are actively in buying mode, so a click on an ad should be rewarded. But with marketing, it's all about building awareness.

  5. DG T from Viewthrough Measurement Consortium, February 22, 2014 at 7:14 p.m.

    Glad to see this article and the IAB stepping up to define viewthrough in their recent white paper on metrics.

  6. James Keyes from RTB International, March 6, 2014 at 12:01 p.m.

    Great article. You take a complicated subject and make it easy to follow and explain to a client.

    I provide big business RTB technology with matching sales funnel design to small businesses and your explanation is a big help to me. thanks!

Next story loading loading..