- Why emotion-based advertising matters and how it leads to greater sales lift.
- The advanced tools available to deliver the right ads and rid ads of bad environments.
- How deep learning models can optimize your ads.
During its run, AMC’s “Mad Men” captured viewers’ imaginations by vividly depicting how advertisers have generated emotional connections to create a stronger bond between consumers and brands for decades. The concept of creating an emotional bond still holds true today, but with the rise of digital, the way brands reach consumers has been refined, opening the door for new tools and approaches.
The advertiser’s enduring challenge remains reaching the right user at the right time. When narrowing in on timing, we usually think of contextual targeting and brand exposure on the preferred platform with placement next to relevant content. However, to be exponentially more effective, advertisers need to reach audiences when the content will resonate the most emotionally. Research has shown that decisions are influenced or determined by emotions. Insinuating emotions can be the driving force behind brand perception, loyalty, and purchase intent. Further, as consumer demand for personalization rises, advertisers that employ an approach to ensure their brand is associated with content that resonates emotionally will achieve a notable edge.
Why are emotions so important to advertising?
In 2016, Nielsen found that ads that generated emotional response saw a 23 percent increase in sales volume, compared to those that lacked emotive elements. This tactic ensures that the ad, which is meant to evoke a particular response, reaches the exact audience for optimal success. For example, a spot for a high-action video game meant to stir up excitement for an impending release would very likely lose effectiveness if placed alongside a video describing the efforts of a charity serving underprivileged children in war-torn countries.
Today, we have more advanced tools to mitigate wrong ads from being served to the wrong audience. This includes sentiment analysis, contextual targeting, and AI tools that can programmatically sift through user data such as their watching, listening and reading behaviors over time. The content surrounding the ad is also analyzed through automatic labeling of non-brand safe messaging of all media content, text as well as image and video meta-data, on webpages, to offer powerful bespoke real-time experiences.
How to do it right
As the digital ecosystem shifts to more transparent data practices, there has been a debate about whether emotion-based advertising can still be achieved. Publishers such as The New York Times, The Daily Beast and USA Today are implementing different tactics to appropriately, and responsibly, cater the right ads with the right content. The New York Times, for example, has been using online surveys that prompt users to select from a multiple-choice list of emotions to glean insight into how they felt after reading an article. Ads on The New York Times saw a 40 percent increase in performance as a result of leveraging this data.
Currently, deep learning models can automatically generate descriptive captions for images, which can be extended to annotate videos on a frame-level. When combined with audio transcripts and text-based emotion classifiers, this tactic can produce deep emotional labels for videos. In order to set this up and deploy successfully requires careful calibration of the training sets and management of the trade-offs. To begin, one must create a granular set of labels that cover the gamut for each emotion, for example from “downright despondent” to “deliriously happy” or “cautiously fearful” to “daringly bold.” Then, bring together an equivalent amount of meticulously labeled data spanning the emotive spectrum in order to accurately train models that can classify across these fine-grained labels. Achieving the right balance can be complex but the results have proven to be both powerful and lucrative.
Consumers want a more personalized advertising experience in conjunction with transparency on how their information is being stored and used. As long as brands have adequate checks and balances in place when targeting audiences, emotion-based advertising can help foster an emotional connection between consumers and brands while increasing ROI.
Latest posts by Yeshwanth Srinivasan (see all)
- Deep Learning Models and Emotion-Based Advertising - January 31, 2020