Motivational-Based Targeting Using 265 Psychological Traits Comes To Marketing

Innovators have begun to find ways to reach consumers without tapping into cookies or personal data by digging into the minds of consumers to determine how people really think. This would enable the prediction of human motivations through advertising, based on the contextual content on a publisher’s web page.

On Monday, Research Measurement Technologies (RMT) and German-based Semasio will announce co-developed technology that understands the motivations of 276 million people in the United States by analyzing 265 psychological traits, and consolidating them into 15 motivational segments that they call “clusters.”

The technology identifies the most important keywords on the page and determines the motivations before serving an ad.

Bill Harvey, chairman of RMT, said the idea is to improve ad targeting through 15 motivations with familiar traits such as security, longing, love and power. These are terms Maslow would have used to build the Hierarchy of Needs.



Initial tests were done with set-top-box data before the two companies combined forces and moved their project online.

For now, the technology is available in the U.S., but there is a plan to expand the offering worldwide.

Kasper Skou, CEO and co-founder of Semasio, said the technology looks for significant words like "anxiety, heroism and childproof" and disregards common words like “she, he, and it.” 

"If the meaning of a word changes in time, the language model changes based on the consumption of the content on the page," Skou said. "The technology learns from current language use."

Semasio uses natural-language processing to analyze the content of web pages and identify its most significant terms and phrases.

The meaning of each page is captured as a weighted keyword cloud and is dynamically aggregated into a Semantic User Profile of the person visiting the pages. It’s all done in real-time.

By training the Semasio artificial intelligence system to turn these word clouds into scores on each of the 15 RMT Motivations, the two companies can identify websites that will provide contextual meaning to specific ads and reach audiences that resonate with these ads.

For digital video, RMT’s and Semasio’s technology can integrate with several types of addressable TV platforms. Two motivations used frequently since the beginning of COVID include Actualization and Security.

RMT, Semasio and Reset Digital now offer the ability to pick specific program contexts, which provides even more lift than picking Resonant networks. Typically, the program level bids are 10% to 15% higher, per the companies.

Skou said the companies also have been trialing “a slightly more complex approach” called cookieless extension, which doesn’t require a cookie or motivational type for the user when they use a web browser.

Context Resonance and ID Resonance can be used separately or together.

For advertisers interested in reaching specific Motivation target audiences, the RMT and Semasio 15 Motivation audiences are available in all leading DSPs, in addition to the LiveRamp Data Marketplace.

5 comments about "Motivational-Based Targeting Using 265 Psychological Traits Comes To Marketing".
Check to receive email when comments are posted.
  1. David Adelman from OCD Media, August 3, 2020 at 11:56 a.m.

    For what it's worth, Charles, any media planners worth their salt have been using motivations to build-out target segments and execute media buys for many years. It's been the starting point for everything we do at my media agency and has been available to us all along. Our approach has not been to create cohort groups as one-size-fits-all for every category. We believe that each category has its own unique sets of motivations and deeply rooted attitudinal beliefs that define how people behave. 

    Just because it's new for the digital world doesn't mean it's new to the world.

  2. Ed Papazian from Media Dynamics Inc, August 3, 2020 at 12:21 p.m.

    Correct, Dave. Consumer mindset  segmentation has been around since the early 1960s and is routinely a critical variable in brand positioning and creative strategy decisions. As for media, when brands are allowed to develop their own media plans---and execute them---in contrast to corporate upfront national TV time buying---it has long been possible to use sources such as MRI, now merged with Simmons, to profile media audiences to get a better fit between the miindset of the viewer or reader and the mindset being cultivated by the ad message. As you said, "Just because it's new for the digital world doesn't mean it's new to the world".

  3. David Adelman from OCD Media replied, August 3, 2020 at 12:26 p.m.

    Thanks, Ed. This isn't the first time digital specialists think they invented something new. Case in point: native advertising is just the digital version of advertorials or product integrations. 

  4. Dave Hills from Spectrum Media Services, August 3, 2020 at 12:58 p.m.

    Great innovation by great companies and pushes the ball downfield for sure. Agree going past individual keywords to understand a page is table stakes going forward. And i'd imagine that resolving to a motivational standard taxonomy is far better than resolving back to individual targeting co's taxonomies that were set up around the company and not the consumer. It's still a standardized way to categorize which we found wasn't as effective as creating plans of segments for brands that are custom to their businesses when using context as a major signal input. 

    We wound up creating a custom search engine with 18 months consumption data at the page level. From this index we deliver topics to the brands audience segments and then the brands can do what they want with it in their own segmentation.  Forcing it into any standardized taxonomy we found weakens the contextual signal.

    We think the goal has to be to be able to aggregate data across models in a flexible way across all media types and formats and in contextual we concluded that requires a custom approach that is still programmatic, delivering scale and precision.  Congrats to Charles and the other companies.

  5. James Smith from J. R. Smith Group, August 3, 2020 at 2:50 p.m.

    Such a schema might work for some clients.  I'd want to learn a whole lot more about the
    black box methodolgy and see some rigorous, and transparent, test cases over time.
    The term "clusters" was used. Does that mean the dominant multivariate procedure is cluster analysis? 

    I agree with David A. and Dave H. regarding concern over pre-set cohorts or targets, even if they are real-time. Further, natural language processing (NLP) is not error free.

Next story loading loading..