AI-powered ad testing—How Delta and other brands are determining the precise impact of their marketing

A bevy of the world’s biggest brands including Delta Air Lines are turning to AI to predict how ads will perform, how the public will react and how much money campaigns are likely to generate before they even launch ahead of some of the biggest marketing moments such as the Super Bowl and the Olympics. In one instance, a popular beer brand, which Ad Age was not authorized to name because of sensitivities around its strategy, used AI to pre-test the emotional resonance of its Super Bowl campaign this year. The beer brand used Daivid, a platform that developed AI to analyze how media is likely to land with the public. Through the AI analysis, the beer maker discovered a key insight—one of its Instagram ads had a stronger effect on consumers’ purchasing intent than even the multimillion-dollar national spot, according to Peter Daboll, Daivid’s head of U.S. That information ultimately helped the beer brand figure out its social media plan around the game, Daboll said. Watch all of this year’s Super Bowl commercials In another example, a global snack brand, which also had its name hidden for this story, recently used the same AI technology from Daivid to understand the sales impact of different creative. The snack maker found that when its ads elicited a sense of excitement from viewers, they were more likely to drive sales, Daboll said. Daivid’s tool conducts detailed emotional analysis, measuring up to 39 different sentiments. “These are miracle insights,” Daboll told Ad Age, and they are changing the way brands take campaigns to market. A brand can relay the AI’s analysis to a creative agency, informing the strategy going forward. “We’re going to look back at the old way of testing ads as the biggest, most inefficient way of gathering data,” Daboll said. NextGen Marketing Summit Join us for a crash course on Gen Z, DTC marketing and the creator economy Register today Delta used similar AI technology to study how much revenue it could attribute to its 2024 Summer Olympics marketing campaign. Delta received daily AI-generated reports that tracked tentpole marketing moments such as when it revealed Team USA signage on planes headed for Paris in May. Daivid is just one of the many upstart and established advertising services turning to large language models to help brands get better information about their ads and make more accurate predictions about the campaigns before they launch. Ad Age also spoke with AI marketing and analytics platform Alembic and ad agency Wpromote, which are developing similar AI tools. The AI models are becoming more adept at predicting consumer perceptions and analyzing brands’ impact than previous approaches, advertising experts said. And as the rise of generative AI enables brands to produce more marketing material than ever before, the risk of wasteful spending is also higher, they told Ad Age, which has placed a new level of importance on knowing precisely which materials will make a difference and which will not. Daivid was unable to disclose the brands’ names for this story, but Nike, Snapchat and DirecTV have also used the platform’s technology. Daivid’s AI tool is trained on data that collectively allows it to determine how an ad may be perceived. Eye-tracking information enables the model to know which visual details in an ad are most likely to attract a viewer’s attention. Facial coding data helps the model understand how facial expressions reflected in an ad affect a viewer. The AI also uses consumer survey data to analyze brand recall and purchasing intent. The AI can help identify that “one thing that you can tell your agency to fix,” Daboll said. More news: AI search platforms characterize brands differently Daivid’s beer client altered the rotation of its Super Bowl advertising on social media after homing in on the one Instagram ad that created strong purchasing intent among the audience. The brand prioritized the higher-performing placement, Daboll said. While the brand also realized that the ad reflected stronger purchasing intent than its actual Super Bowl spot, the insight was gleaned too close to the game to make any changes, Daboll said. The global snack brand that found excitement was the ideal emotion to summon in its ads used that to inform its creative strategy going forward, Daboll said. Wpromote developed an AI ad testing tool called Creative Audit, which uses 72 different criteria to compile a score for how well an ad helps a brand meet specific objectives, such as brand lift, driving purchases and increasing website traffic. The factors that make a difference can be subtle, said Michael Stone, the agency’s chief relationship officer. Whether an ad starts with energetic momentum, for example, is one factor the AI looks for; another is the degree to which human faces are included in the ad, or the prevalence of a brand’s logo. Creative Audit can be used for social media to predict the social video creative that is most likely to perform best on TikTok, Instagram and YouTube, according to Stone. Apparel brands, especially, are using Creative Audit, Stone said. The tool is still nascent, and Stone declined to name the brands that are testing it so far. Given the speed of the scroll on social media, it’s important to be as precise as possible with the creative to capture the fast-moving attention of viewers. Creative Audit is meant to help brands determine likely performance before the ads even run. Delta’s use case Brands including Delta are also using the predictive power of AI for their entire marketing strategies. Delta tested ad tech platform Alembic’s AI system for its large presence around last summer’s Olympic Games, which included sponsoring Team USA’s travel, hosting on-the-ground programming and running TV commercials. The use of AI is reshaping how Delta measures the impact of such official sponsorships, according to Jake Abel, head of marketing operations, strategy and media for Delta. “As a brand marketer, you typically don’t get a lot of insight about revenue and sales [from these events],” Abel said. Alembic’s AI model was able to scour vast amounts of data collected via Delta’s marketing and return daily and weekly reports about what was driving sales and revenue goals and what was not. These kinds of performance reports would have previously taken far longer to return to Delta, Abel said. For example, Alembic showed that Delta’s promotions around its Team USA-themed aircraft livery—such as a widely-viewed segment on “Today”—had a particularly strong effect on revenue by driving clicks to Delta’s website. This insight would have been too in-the-weeds for any human marketer to surface, Abel said, especially given the plethora of other data points obscuring causal inference—such as how Delta’s TV spots and sponsoring of on-the-ground programming was affecting consumer perceptions and their purchasing. Moreover, manually connecting these data points to find key drivers of success is a more costly approach than using AI because of the greater time required to do so, Abel said. View this post on Instagram A post shared by Delta Air Lines (@delta) Alembic’s AI also determined that Delta received a bump in its sales goals when fellow Olympic sponsors mentioned the brand. The finding might seem obvious, but quantifying it pre-AI was difficult, Abel said. The insight was also a stark reminder that building awareness is still the fundamental task of brand marketing, Abel said. Recent news: The top 5 creative campaigns to know about Delta did not alter any of its Olympics marketing based on Alembic’s reports because the test was meant to be a proof of concept, Abel said. But Delta is now in the process of standing up Alembic’s system so that it can be used to make changes to its brand marketing strategy, based on evidence of what is driving surges in sales—from TV spots to promotions of partnerships and other sponsorships. “The more that marketers’ ‘gut feeling’ can be backed up by science and data, the better,” Abel said.

AI-powered ad testing—How Delta and other brands are determining the precise impact of their marketing

Featured Stories