DTC Briefing: How skin-care brand Geologie used AI to redesign and package its acne treatment line

In this case, Rehab AI used first-party data gathered from Geologie’s own customers as well as the brand’s clinical study to create 20 detailed customer personas ranging from “budget-savvy teen” to “career-focused parent.” Geologie could then feed different information — including sample taglines and sample ad copy — into custom AI agents, and Rehab AI would spit out different feedback that each of these personas would be expected to give. Geologie used this process to develop its new line of packaging, which is rolling out on Target shelves throughout this month.

It’s just one example of how direct-to-consumer brands are increasingly looking to AI to streamline their operations. When ChatGPT first burst onto the scene, many direct-to-consumer executives started by spending a couple of hours a week experimenting with these new AI tools, seeing what answers they would provide if asked to come up with names or new product lines, or to summarize meeting notes. But now, more brands are leveraging AI to streamline or completely replace tasks, like orchestrating focus groups, which have been staples for decades.

It remains to be seen exactly just how well AI does at replacing some of these tasks. But Geologie said early customer feedback on its new packaging has been positive. What’s more, “We spent a total of 10-12 hours, with the use of AI, to accomplish the same results that could have taken up to 85 hours,” Stephen Racano, head of growth at Geologie, said.

Integrating machine learning and generative AI 

Geologie, founded in 2018, sells products like cleansers, moisturizers and pimple patches. Its products are sold through its direct-to-consumer website, on Amazon and at Target, where it launched in February 2024 and is currently in nearly 1,000 doors.

Racano said the company first started dabbling in AI between 12-16 months ago, by way of ChatGPT. “It took months to learn prompts and teach ChatGPT about the brand and how to talk about Geologie,” he said.

“People are increasingly putting together skin-care regimens using AI chatbots,” Racano said. In turn, Geologie has seen increased traffic from ChatGPT and other AI search bots. Geologie has invested more consistently in its SEO content, and Racano said Geologie’s blog posts regularly show up on the first page of Google for queries like “oily nose” or “why is my face oily when I wake up?”

Last October, when it came time to redevelop the new Clear system of products for Target, the company decided to delve deeper into how it could use AI to more efficiently make sense of its first-party data to save time during the product development phase.

When Geologie launched its Clear acne treatment system at Target last year, it bundled all of the products, ranging from a night cream to a clarifying gel cleanser, into one kit that customers bought together. But, as part of the refresh, Geologie decided to break up the set into individual products and offer each of the skin-care products in a larger format and with a new packaging design. And that’s where Rehab AI came in — to help determine how to make each product stand out on shelves while clearly and instantly communicating to customers the use case of each product.

The AI model was fed thousands of reviews mentioning Geologie products, Racano said. “We also scraped a bunch of Reddit threads and brought all that data in to create our personas,” he said. “Previously, I would read all those Reddit threads and write a blog article about them, which took forever.” 

Another set of data that was fed into the model was one of Geologie’s biggest clinical studies, released a year ago, which made a case against the use of benzoyl peroxide as an acne-fighting ingredient.

Geologie’s formulations instead use ingredients including salicylic acid, azelaic acid and hyaluronic acid, among others. The company wanted to leverage the benefits of its ingredients in the packaging and messaging, Racano said, “which brought us to our new tagline: ‘Bad for acne, good for you.’” 

Once Geologie fed all this data into Rehab AI’s custom AI agent, it then asked the bot to provide feedback on what different customer personas may think of also messaging that the products were vegan or cruelty-free. Racano said the AI bot suggested Geologie instead mention that its products are free of parabens, as an example. “Along with that feedback, [we considered data collected] from a year of being on Target shelves,” he said.

In addition, the AI suggested that Geologie make its products’ labels larger. “The AI knew customers would be standing away from the shelves, so we adjusted the labels a few times during the redesign process,” Racano said.

The importance of first-party data

Racano said Geologie wants to be at the forefront of AI use in product development, not just marketing material creation.

Namrata Shah, a partner and global head of consumer and beauty at PERLab, the product redesign practice at Kearney, said AI models are a new way for CPG brands to revamp their R&D strategy.

“Most CPGs, including big beauty brands and conglomerates, have not evolved their innovation processes in the last five to 10 years,” Shah said. Many brands still follow the traditional stage-gate process, which can take up to 24 months from idea to shelf, through a manual approach with a lot of inefficiencies. Shah said this is often due to outdated linear processes that involve multiple hand-offs across siloed teams, as well as bureaucratic models and repetition.

“I see a pivotal moment for the beauty industry to leverage AI to cut time and cost of innovation by half and to double the ROI,” Shah said. This can support accelerated product testing and validation by leveraging existing combinations of ingredients in the brand’s portfolio, among other rapid development processes. “Synthetic personas will allow us to autonomously stress test within minutes,” Shah said. That could include testing what formulas consumers care about, what marketing claims will resonate and what customers are willing to pay for a product.

“The packaging design process is definitely one area that we’ll continue to see AI help accelerate,” Racano said. However, he said, it is important for brands to bring in first-party data to help sufficiently inform any custom AI bots about what the company stands for. “Otherwise, it’s a very generic use of AI against competitors who are using the same tools,” he said. –Gabriela Barkho.

By the numbers: Hims & Hers earnings

Hims & Hers is one of the DTC brands that has been on a tear over the past year, consistently raising its revenue forecast. That’s thanks, in large part, to its decision last year to start offering prescriptions of compounded semaglutide, a custom-made alternative to brand-name GLP-1 drugs.

This product provided a boon to Hims & Hers at a time when drugs like Ozempic and Wegovy were in short supply. But, last week, the FDA declared that the semiglutide shortage was over.

Hims & Hers tried to put a positive spin on the news. “Now that the FDA has determined the drug shortage for semaglutide has been resolved, we will continue to offer access to personalized treatments as allowed by law to meet patient needs,” Hims & Hers CEO Andrew Dudum said in a social media post on Friday. And, when Hims & Hers reported its fourth-quarter earnings on Monday, it once again raised its full-year revenue guidance. Here’s a closer look at how Hims & Hers did during its most recent quarter, by the numbers. –Anna Hensel

$481 million: Fourth-quarter revenue reported by Hims & Hers.
77%: Gross margin
2 million+: Number of subscribers on the Hims & Hers platform.
43%: Full-year revenue growth excluding contributions from compounded GLP-1 drugs.
63%: Year-over-year increase in AOV during the fourth quarter

Quote of the week

“What I think everybody’s realized in DTC is that chasing growth and then sort of figuring out how to make money and make the business model work at a later date is a really challenging way to build a business. … I’ve seen the right way to build a business, and it’s a combination of brand, culture and smart cash management.”

–Jill Layfield, CEO of bridesmaid dress company Birdy Grey, on how the DTC world has changed over her two-decade retail career spent at companies including James Michelle, Tamara Mellon and Backcountry –Julia Waldow

What we’re reading

• Solo Brands CEO Christopher Metz, who took on the job just last year, is exiting the company. 
• Bloomberg has a deep dive into how Walmart “wants to be something for everyone in a divided America.”
• Shein’s profit dropped by almost 40% last year, according to the Financial Times, citing anonymous sources.

What we’ve covered

• Once relegated to farmer’s markets and natural grocers, aluminum-free deodorant brands like Curie and Native are now taking over mass retail. 
• Vita Coco took a DIY “vending machine” — made of cardboard and a utility cart — to New York City’s Washington Square Park to hand out samples in a cheeky nod to Poppi’s Super Bowl stunt.
• How the rise of GLP-1 drugs is upending inventory planning.