March 23, 2023
Hill & Markes was founded in 1906 as a family-owned candy, ice cream cone and ice cream distributor in upstate New York. Today the company is known for a plethora of products, such as janitorial supplies, office supplies and food service disposables. While the product lines expanded and changed over the years, the seasonal nature of its business did not. Hill & Markes sought a simple solution to increase ecommerce revenue with a quick and easy-to-implement solution.
The goal for Hill & Markes has always been to generate more revenue from their website. Their challenge was getting customers to find and purchase new and relevant items, explained Nick Seefeld, ecommerce manager.
Seefeld manually coded product recommendations for the ecommerce site to enhance discoverability. While Seefeld was intrigued with automating this task with artificial intelligence (AI), he’d never experimented with AI and had a lot of uncertainty about AI solutions.
The Hill & Markes team was also concerned about its data and the potential challenges of integrating new technology into the company’s existing systems. Seefeld didn’t have time to spend on development to ensure integration was seamless for all 50,000 products sold on their website.
Hill & Markes worked with Proton to implement a product recommendation AI model.
“In terms of our technology vendors, Proton has far and away been the easiest to work with,” said Seefeld. “We had regularly scheduled meetings, constant conversation and collaboration on the project. They made it seamless and painless. They finished the integration ahead of the original five-week schedule they proposed.”
Once Hill & Markes gave Proton access to their site and shared their data, the Proton team handled the integration, allowing Seefeld to focus on other objectives. Now, when customers visit the Hill & Markes ecommerce site, there’s a rotating panel of products toward the top, titled “Recommended for You.”
The AI tool pairs nicely with the other integrations already on Hill & Markes ecommerce site, and Seefeld confirmed that the website is delivering relevant product recommendations to customers— no surprises or mismatched suggestions.
Proton’s solution not only embeds product recommendations, but also tracks when customers click on those recommendations and how people browse, which is data distributors without this technology struggle to capture.
An easy integration with Proton meant quick results for Hill & Markes. During their A/B test, Hill & Markes saw a 43% increase in average revenue per customer with Proton’s AI-powered product recommendations. Since then, total web revenue has continued to climb. August and September of 2021 had the highest web revenue on record. September 2021 ecommerce sales were 53% larger than September 2019. Additionally, the average order volume at Hill & Markes typically dips in August and September, but didn’t in 2021 with Proton in place.
For other distributors considering integrating AI but are hesitant about adopting new technology, Seefeld recommends doing an A/B test to see the results.
“The results of the A/B test blew our expectations out of the water. I was thinking that maybe during the first quarter we’d be ramped up, but this increase was instantaneous,” said Seefeld. “And the entire process was painless.”
.svg)
.svg)
.svg)
