Imagine that you’re the CEO of a major beauty retailer. You’ve released an augmented reality app that lets customers test makeup colors, even out their skin tone, or try out lash extensions—like using an Instagram filter, but with products from your store. It’s a huge hit, and sales are going gangbusters.
But then something in the user data catches your eye: the most highly engaged users are young girls, and many of them are using the app between 1:00 and 2:00 a.m. Some of the feedback they’ve given the marketing team is disturbing: “I only like myself when I’m using the app.” “I don’t like the real me.”
Though the technology has changed, it’s a dilemma familiar to marketers in this industry. Customers’ insecurities about their appearance are a main driver of beauty-product sales; if engaged ethically, these girls could become loyal customers for decades to come. But nobody wants to exploit a vulnerable population or inadvertently lead them to harm.
These quandaries are becoming even more salient as AI and big data offer marketers new ways of engaging customers through increasingly powerful personalized marketing. For this particular retailer, shutting down the app would shave off 20 percent of the company’s valuation overnight. Keeping it live will clearly take a toll on some young users’ self-esteem. What should the company do?
Establish monitoring mechanisms
In the case of the beauty company, their promising new app could quickly become a crisis. Voices within the firm may be concerned enough about the impact of the campaign on vulnerable groups to want to see it discontinued. The company may also have concerns about external criticism from customers or advocacy groups. That puts the CEO in a difficult position, and there’s no single right answer for how to proceed.
Sawhney recommends that the best form of crisis management for any company is preventing the situation from rising to the level of a crisis in the first place. This starts with proactively establishing robust guardrails and monitoring mechanisms to flag unusual or worrying activity by customers.
In the case of the beauty-product company, systemic monitoring would have warned that young girls were showing higher-than-average engagement with the AR tool, or that sales figures showed they were buying expensive acid-based wrinkle creams intended for mature consumers that could actually harm delicate young skin.
“Why would you not have any early warning flags?” Sawhney says. “It’s a well-known fact that the preteen demographic is vulnerable.”
If a CEO does catch something that they’re not comfortable with—or that they think their customers won’t like—their best bet is to act swiftly and honestly.
“You have to make a dispassionate decision on how to contain the blast radius,” Sawhney said. “Make an apology, and make it quick. Don’t hide it. Don’t pretend it doesn’t exist. Covering it up is worse.”
For example, when Meta realized it had an issue with younger users, it eventually created Instagram Teen accounts with parental controls and other safety measures. But that only happened after its founder, Mark Zuckerberg, struggled to explain to Congress why Meta had been ignoring research showing that Instagram harmed teen girls’ self-esteem.
“You want to get ahead of it and say, ‘Hey, we’ve looked at the data. We don’t like what we see, and even though we’re not required by the regulators to do anything, we are taking proactive measures,’” Sawhney says.
For the beauty CEO, that could mean shuttering the app at certain hours to ensure it’s not depriving school-age children of much-needed sleep or distracting students when they should be focusing on class.
“You can mitigate these problems if you are proactive,” Sawhney says.
Develop positive campaigns
Today’s marketers have tremendous power at their fingertips—from sophisticated psychological research to data mining and AI-powered personalization. Sawhney recommends bringing these tools to bear in developing positive marketing strategies for engaging with customers in vulnerable groups.
Sawhney points to two well-known ad campaigns that each began from a place of empowerment. The “Like a Girl” campaign from Always originated when company executives connected the pivotal moment when girls begin using feminine products to all of the other changes that happen, for better or worse, during puberty.
The marketing team found that during puberty, girls’ self-esteem drops, they stop playing sports, and they get socialized out of STEM classes. The company countered that data by producing videos that showed young women running and throwing with confidence and strength, describing in positive and moving terms what it meant to do something “like a girl.”
The personal care brand Dove has also leaned into this trend through slogans including “Let’s Change Beauty” and “Real Beauty.” Its “You’re More Beautiful Than You Think” campaign featured a video where a forensic artist draws women’s faces using their own descriptions of their features, then again based on descriptions of them provided by their friends. The images, side by side, show that women perceive themselves much more harshly than others do.
“These campaigns proactively shift the brand position to one that stands for empowerment,” Sawhney says.
Get social, but not too personal
Companies with engaged communities on social media can mine those conversations to detect pain points or other opportunities for empowering messaging. For example, analyzing the discussions of cosmetics shoppers might reveal that some young women are being bullied at school. That could give the beauty company a chance to reframe beauty products as tools for self-expression and a means for reclaiming agency and creativity.
Social media may also be a good place to observe how customers are using tools like the augmented reality app, both for monitoring concerning uses and finding new ways to improve the technology. But be careful about prying too close into customers’ lives and preferences, Sawhney says.
As AI becomes more integrated into marketing tactics, decision-makers will find themselves walking a delicate tightrope between being helpful and, well, creepy. Customers appreciate it when firms give them relevant products and messaging. But that very personalization can cross boundaries of consent. Groups that already feel marginalized or vulnerable may react negatively to companies that seem to know too much about them or use their data too aggressively.
“AI can facilitate much more meaningful and relevant engagement with your customers,” Sawhney says. “But just because you can personalize to the nth degree, doesn’t mean you should.”
In the end, marketing ethically calls for new ways of thinking that go beyond pitting the interests of people against the need for profit. That’s especially true when it comes to sensitive or vulnerable groups such as young people whose brains and senses of self are still developing. Sawhney acknowledges that this is easier said than done, especially in the age of AI.
“I think we’ll have to be a lot more creative to come up with ideas and interventions that are good for vulnerable demographics but also make good business sense,” he says.
