DoubleClick Experiment Could 'Re-imagine' IAB Display Metrics

Google said it will run a DoubleClick for Advertisers experiment, with help from brand measurement research firm Vizu, that could change the way companies measure brand lift from display ads.

The company will present the findings to the Interactive Advertising Bureau.

DFA experiments began in mid-2008, but this could be the first with findings presented to the IAB that change brand lift measurements. The pilot aims to show that the use of experimental design principles -- a research term -- is a good way to address brand measurement challenges and to provide actionable insights to brand marketers.

Sherrill Mane, SVP of research, analytics, and measurement at the IAB, notes that if successful,the research could have far-reaching implications for the future of brand impact measurement of online advertising campaigns.

Surveys are a measurement tool already used by marketers to measure the effectiveness of a campaign, but, there are some challenges today with experimental design and how the responses are collected. "We hope to move the industry forward by establishing a true experimental design form for this measurement," said Sanaz Ahari, senior product manager at Google. "Online brand measurement for display is a pretty significant investment for the industry, and the IAB has been talking a lot about a solution around true experimental design, which is our focus." 

Through the research, the team will conduct online brand measurement research for about 70 campaigns using DoubleClick for Advertisers. Google will use DoubleClick ad-serving technology to create two randomly selected user groups -- one will see the campaign's ads; and another, public service ads. The experiment will tap Vizu's sampling technology to serve a brand lift survey to all users. The study will individually measure variables like brand awareness, purchase intent and favorability. The pilot will run between three and five months.

This data can be used with the other DFA campaign metrics, such as reach, media cost, conversions and attribution, clicks, helping marketers to see the impact of the campaign using any metric they want -- from impressions to clicks to brand impact to conversions.

While Vizu uses true experimental design in its methodology, Scott Schwanbeck, VP of business development at Vizu, said the company always looks at ways to improve industry metrics and drive online brand research. Google will cover the cost for the research.

A previous Google DFA Experiment measures conversions. Some 80% of campaigns showed increased conversions as a result of display ads.

Earlier this month at MediaPost OMMA Global in San Francisco, research firm Brand Keys released 2012 findings from its annual Digital Platform Engagement Index, which analyzes how digital changes branding for companies. While it does examine the influence of true experimental design on brand metrics, the study finds that it becomes less important to analyze the audience demographics, such as age and gender, and more important to learn more about how 14 digital platforms -- such as display, social and search -- intersect with product and service categories, as well as what consumers most want to talk about and where.

5 comments about "DoubleClick Experiment Could 'Re-imagine' IAB Display Metrics".
Check to receive email when comments are posted.
  1. Tara I from Google, March 29, 2012 at 10:02 a.m.

    Who can name 14 digital platforms? This number seems high ...

  2. Tara I from Google, March 29, 2012 at 10:20 a.m.

    I guess if you count the providers of the platforms ... bad sentence.

  3. Jimmy Bogroff from Emerge Interactive, March 29, 2012 at 4:07 p.m.

    Ironically,Style has roadblocked this page and wants to make sure that I'm aware of their position as the fastest growing cable network among women 18-49.

  4. Jimmy Bogroff from Emerge Interactive, March 29, 2012 at 4:14 p.m.

    oops wrong comment / wrong article!

  5. Ray Pettit from PRN, March 29, 2012 at 5:05 p.m.

    "Google will use DoubleClick ad-serving technology to create two randomly selected user groups" In 'true' experimental design, it is random assignment, not selection, that creates two probabilistically equal groups. I'm sure that is what Google is doing...

Next story loading loading..