CampaignTester™ Press

AI Emotion Analysis Would Have Predicted Responses to the Peloton Ad

Lost Stock Value & Holiday Sales Can Be Prevented by Testing Ads for Emotion Before Investing Millions in TV & Digital Campaigns

Shelli Garson
December 9th, 2019

“What were they thinking?” is the question asked – or the underlying thought behind many of the comments shared on social media about the Peloton TV ad. They should have known better-  because they could have known better. Emerging artificial intelligence platforms use machine learning to detect emotion, sentiment, and engagement of potential customers or other important stakeholders. While there are still very good uses for traditional focus groups these newer technologies allow very fast testing at extremely low costs – compared to the traditional focus group methodologies. Marketers and brands now have no excuses not to test before spending millions – only to get surprised by negative reactions from their best potential customers – or voters –  in the case of 2020 political campaigns

While Peloton is the latest brand to “miss the mark” with its target buyers, there are many others. as Suzanne Vranica writes, 

“Brands including Nike Inc., Proctor and Gamble Co’s Gillette and PepsiCo Inc. have released ads that stirred controversy in recent years,” writes Suzanne Vranica, Advertising Editor of The Wall Street Journal, “Peloton Thrown by TV Ad Backlash.” (link

Comedians, late-night talk shows and others have made a living creating parodies of these major branding mistakes. This week Saturday Night Live (link) and Ryan Reynolds (link) created great comedy by simply “flipping” the sentiment and emotional reactions of the audience. Brands can prevent the loss of brand equity and becoming the subject of comedy and parody by pre-testing sentiment and emotion. In addition to “brand protection,” businesses and brands are increasing investment into sentiment analysis into their suite of more traditional business intelligence for a variety of uses beyond marketing. In fact, according to Deloitte, “Emotion-driven Engagement: Foster Deeper Connections Across Key Moments with Customers” (link)  the investment by businesses in sentiment data is predicted to grow into billions globally.

Would emotion detection technology mirror the emotions expressed through social media? I decided to put the CampaignTester (link) platform through a “test” of the Peloton ad to see how the audience reacted to the commercial. 

CampaignTester uses the front-facing camera in a mobile device (after getting a person’s permission) to pre-test video and TV content. The analysis of micro gestures and other very subtle changes in facial geometry are compared to millions of other reference images of known emotions in the company’s proprietary data science and process. The reports provide a second-by-second emotional analysis of video content that tracks eight different emotions. The company also tracks “engagement” which in this case means that a person is facing the screen and that their eyes are on the screen. The intuitive CampaignTester data dashboard allows marketers to see the specific thumbnail frames of the video as a specific reference for reactions within the videos.

Here is an overview of the Peloton Ad results:

Using CampaignTester and its viewer community, this :30 commercial was tested among an audience of 100 women and 80 men using the mobile app on their phone. The test enjoyed a strong activation rate, where nearly all respondents were compliant within a few hours of the test. The data proved both fascinating and predictable and echoed the verbal sentiment expressed in social media.

The test uncovered that indeed, the messaging of this commercial, while well-intended was not received as it was meant to be.  Designed to reflect on “joy”,“gift-giving”, “Christmas”, and “surprise” it appeared to alienate viewers more than attract them, with the emotions of confusion and sadness running high throughout the commercial, sprinkled with a touch of anger (higher among women.) That means that the lion’s share of emotions measured in this study are more likely to express negative emotions than positive ones. 

Strikingly absent from the range of emotions expressed was the sentiment of surprise and joy. Was this not meant to be the tone of this emotional message? Clearly, this does not bode well for the brand nor for a Christmas gifting message.

Further, we observed a spike in confusion at the end of the commercial, indicating a lack of a rich and compelling selling proposition and “call-to-action: among respondents.  Not the happy ending marketers want for their brand!

A similar video test using the CampaignTester platform would have cost under $1,000. The pricing is based upon the length of the video and the number of people tested. There is also an option to add gift cards as an incentive for people to watch the videos as testers – similar to some of the leading survey research companies.

“Direct Brands” like Peloton and a growing number of other companies need to be especially cautious with messages because they have direct communication with their customers and stakeholders. Investing in sentiment and emotional analytics has never been more critical given the disruption media can stage in an instant. Using these data can help make sure that marketers get the messaging “right” and can change a “flop” into a “hit” by making changes in advance of investing time and money into paid, organic and social channels.

Politicians and political campaigns would be well advised to consider sentiment analysis for their creative strategy. There are many complexities in issue-based messages that resonate with different geographic and demographic groups of voters. CampaignTester was one of the companies that showcased their technology at the Campaigns and Elections Innovation Summit last month in Washington, DC. (link) Sentiment analysis allows the tuning and refinement of messages that could be the difference in winning elections.

As a researcher, the growing science around AI, micro gestures, machine learning, and emotional reactions make utilizing this new data a “no brainer” and an absolutely essential component of any video campaign for many marketers. Video testing platforms like CampaignTester are very affordable and may help avoid volatility in stock prices, severely impact sales and damage brand equity.

And that’s my unemotional assessment of using emotion and engagement data for your business or brand!

About Shelli Garson:

Shelli Garson is a seasoned marketing researcher and focus group moderator with over 30 years in the industry. She has delivered market research and brand strategy for more than 100 brands including AutoNation, McDonald’s, Mattel Toys, Bacardi, Dunkin, Art Basel, Wyndham, Christian Dior, Rockport Shoes, AMF, Air France and others.

source

By |2020-02-27T15:03:42-05:00January 16th, 2020|Emotions|Comments Off on AI Emotion Analysis Would Have Predicted Responses to the Peloton Ad