How do AI-generated ad slogans compare to those created by professionals? New University of Minnesota research suggests that the pros tend to connect better with consumers, but AI tools may perform just as well by adopting certain strategies.
Advertising slogans often require nuance, as marketers must consider cultural context, brand fit and emotional response when crafting them. AI tools can generate multiple versions of ad copy in a fraction of the time, but it can be difficult to determine whether their output is effective. Joe Redden, a professor of marketing in the Carlson School of Management, tested this in new research published in the Journal of Current Issues & Research in Advertising.
Redden conducted two studies in which he prompted AI to create multiple alternative taglines for 10 different ads for well-known brands that ran in national magazines. Then, he edited versions of the original ad and had 6,000 consumers rate these newly created ads for likeability and brand purchase intent. In the first study, he used ChatGPT-3.5 with a simple prompt to generate the taglines. In the second study, he used ChatGPT-4 with a more detailed prompt that assumed the role of a creative ad executive, considered the target audience and described the accompanying image in the ad. In each study, he employed a "test for the best" approach to identify the most effective ChatGPT option for the ad based on the consumers' ratings. The study showed:
- On average, ChatGPT taglines underperformed the original ad tagline in nearly every product category. This suggests that marketers should not trust most taglines generated by ChatGPT to perform well.
- The top-performing ChatGPT tagline performed nearly as well as the original tagline, and sometimes even better. This shows that marketers should test among the many alternatives that ChatGPT can quickly generate.
- ChatGPT-4 taglines generated from a detailed prompt outperformed the ChatGPT-3.5 taglines created from a basic prompt.
"Nearly every business person I speak with is struggling to figure out how and when they can benefit from AI tools," said Redden. "The approach of testing multiple AI-generated taglines rivaled the meticulously crafted, professional ads that ran in a national campaign, yet the former required little cost, advertising skill, technical capabilities or special brand insights."
Redden notes it's important that ad effectiveness goes beyond one slogan, as a successful campaign depends on a range of factors. However, moving forward it will be key for agencies to adopt AI as a tool to remain competitive.
"Practitioners should consider when using generative AI will make sense for them," said Redden. "Using AI along with the 'test for best strategy' would seemingly be a good option for any marketer that simply does not have the in-house expertise or budget to pay for an agency to develop their content."
