Leaf is a UK-based challenger brand in the laundry space that sells dry laundry sheets. Not only does it need to get consumers to switch from liquid detergent, it also competes against giants like Procter & Gamble’s Tide.
First, it tried the DTC playbook. The eco-friendly household and commercial cleaning product maker ran paid social ads on Facebook and Instagram through Google Ads to drum up its DTC business. But it wasn’t converting enough customers, according to Calum Hutchison, Leaf’s founder and CEO. “We stepped away from DTC marketing to focus more on the retail side of things.”
In its previous campaigns with retailers, Leaf ran “pretty generic” digital ads that complemented its in-store physical signage and shelf display units, Hutchison said. Still, the brand struggled to measure the impact of digital marketing initiatives on retail store sales.
In December, Leaf turned to a new tactic: AI. The brand tested an automated creative optimization platform for brands called AdSapiens to create and optimize targeted ads for customers. Leaf launched a one-month campaign with one of its retail partners targeting shoppers based on their proximity to the retailer’s 30 locations as well as transactional data.
Officially launched in late April by digital advertising firm Adludio, AdSapiens created three mobile display ads for the campaign. One concentrated on how Leaf’s cruelty-free products were not tested on animals, a second emphasized the products were good for sensitive skin, and a third was a more general ad. Based on the ad content, AdSapiens built audience segments that were likely to respond well to the ads and found publishers that would match these segments. It also optimized the ads in real time throughout the campaign.
The cruelty-free ad performed the best, but engagement across all three ads represented an improvement over Leaf’s previous efforts. The company’s sales immediately increased, Hutchison said. At the end of the month, sales doubled compared to the previous month.
In addition to the higher conversions, Leaf gained insight into its audience. It turned out that Leaf customers skewed young, drove electric cars, were interested in renewable energy and largely lived in London, Birmingham and the Midlands. Based on these audience attributes, the platform identified ad placements on The Guardian, The Mirror, ign.com, the Birmingham Mail, Business Insider and Manchester Evening News that performed well.
Black box begone
Behind the scenes, AdSapiens uses a combination of large language models (LLMs), machine learning algorithms and computer vision to analyze user engagement and adjust campaigns in flight, said Ian Liddicoat, CTO of Adludio. The LLMs train on historical campaign data from thousands of Adludio’s past campaigns, deriving best practices that inform ads. Separately, the LLMs ingest customer data, such as creative guidelines, compliance policies and brand safety and suitability standards, that the business can upload to the platform. That way, the platform is well versed in both the business and campaign, allowing it to construct ad units as well as optimize publishers and audience segments, Liddicoat said.
To construct high-performing ads, the platform deconstructs them. It divides the ad unit into individual pixels and looks at whether ad components drive increased user engagement or not, according to Liddicoat. It examines elements such as copy, use of color, object placement, CTA button size and what Adludio calls interaction type. Interaction type refers to user actions like tapping on an icon, clicking on objects or creating objects.
For Adludio, whose clients include Coca-Cola, Nike, Bobbi Brown, Neutrogena and Audi, it’s important for clients to have a clear read on the data coming in and the insights it generates. “There’s a responsibility to make sure that AI is not a black box and delivers against the promise of digital,” Liddicoat said.
Though the campaign AdSapiens handled for Leaf was fully automated, the brand received frequent updates about how the tech was learning and adapting, deciding which customers to target and converting them to sales, Hutchison said. He’s looking to extend the work with Adludio to more retailers and other areas of Leaf’s business. “While I’m no authority on AI,” he said, “we’ll be continuing to use this type of software whenever we can.”