About

Stone Creek Coffee is an independent specialty coffee company based in Milwaukee, Wisconsin. It roasts its coffee in-house and sells through its ecommerce store, cafés, wholesale partnerships and regional grocery retailers. The company’s online range centers on coffee beans, subscriptions and products designed around different tastes, roast levels and brewing preferences.

Looking for growth from its Shopify store, Stone Creek partnered with Shopify CRO agency Blend Commerce in 2025 to find growth through A/B testing with Intelligems.

Challenge

Stone Creek needed a repeatable, data-driven way to decide what to build next.

Stone Creek was attracting substantial traffic to its website, including customers researching its cafés and people shopping for coffee. The team believed the store should convert more of these visitors into e-commerce customers, but it had no reliable way to test potential changes before committing development resources.

Its existing development support was largely reactive. The team could fix problems and implement requested changes, but Stone Creek didn’t have a way to reliably identify opportunities, form hypotheses, and determine whether changes genuinely improved commercial performance.

“We couldn’t A/B test the changes we wanted to make, and our developer functioned more as a 'break/fix' role, and they were not strategic partners designed to help grow our business,” explains Eric Resch, Managing Director and Founder, Stone Creek Coffee.

This created two related risks. Potentially valuable ideas could remain untested, while plausible changes could be rolled out without knowing whether they helped or harmed conversion.

Stone Creek needed a repeatable way to assess its customer journey, prioritize opportunities, and make decisions using observed customer behavior. The company’s most important measures included orders, conversion rate, average order value, and revenue per visitor, with additional focus on subscription performance.

Solution

Building a continuous experimentation program with Intelligems

Blend began with a CRO audit. The findings were converted into a prioritized testing roadmap covering product pages, collection pages, navigation, search, subscriptions, merchandising and the cart.

Intelligems provided the testing infrastructure needed to move from recommendations to controlled experiments. Stone Creek and Blend established a monthly cycle, from hypothesis formation to variant design, running experiments and assessing them with Intelligems, then rolling out winners or applying learnings.

Blend can operate quickly thanks to the scalability Intelligems brings to testing.

According to Eric Resch, “Many of our A/B tests have been impactful. What stands out to me is the cadence and care Blend brings to the monthly planning and the tests we run each month.”

The program covers both Blend recommendations and ideas proposed by Stone Creek. Tests have included subscription versus one-time purchase defaults, value propositions below the add-to-cart button, express checkout options, payment messaging, product badges, grind and size information, search visibility, collection-page merchandising and mobile navigation.

One of the strongest experiments reconsidered how information was presented on a key product detail page. Stone Creek wanted to test an alternative layout inspired by another coffee brand, with clearer product copy and more visible purchasing information.

The experiment ran across all devices. Two variants were tested against the existing product page.

Variant 1 increased conversion rate by 40% and revenue per visitor by 25%. Subscription revenue per visitor increased by 50%, while subscription orders per visitor rose by 115%. Although average order value fell by 10%, the increase in completed purchases more than compensated. The variant reached statistical significance and was made permanent.

Another test examined the value propositions displayed below the add-to-cart button on product pages. The revised version gave shoppers clearer reasons to buy at the point of purchase.

The variation produced a 10% increase in revenue per visitor, a 5% increase in conversion rate, and a 10% increase in add-to-cart rate. Subscription revenue per visitor increased by 45%. After reaching statistical significance, the new version was rolled out.

The program has also shown the importance of testing experiences separately by device. Blend redesigned Stone Creek’s mobile category menu to make important shopping routes easier to understand and access. Blend tested two versions against the existing navigation.

The winning version increased ecommerce conversion rate by 20%, revenue per visitor by 30%, and average order value by 10%. It also increased subscription revenue per visitor by 5%. Once the result reached statistical significance, Stone Creek implemented the new menu permanently.

Intelligems has also prevented harmful rollouts, consistently allowing Stone Creek to identify negative effects early and retain the better-performing original experiences.

Results

Stone Creek grew conversion rate 40% and revenue per visitor 25% on its top PDP test with Intelligems

Stone Creek’s experimentation program has improved product discovery, navigation, product-page communication, and subscription merchandising.

Among the results recorded through Intelligems:

• A redesigned PDP increased conversion rate by 40% and revenue per visitor by 25%

• A new mobile category menu increased conversion rate by 20% and revenue per visitor by 30%

• Stronger value propositions below the add-to-cart button increased revenue per visitor by 10%

• Removing express checkout options from the PDP increased revenue per visitor by 20% and conversion rate by 15%

• A revised mobile collection experience increased revenue per visitor by 10% and subscription revenue per visitor by 385% on mobile

With Intelligems, Stone Creek can now test its own ideas alongside opportunities identified by Blend, learn from unsuccessful variants, and invest development effort in experiences supported by customer data. “The impact has been significant,” notes Eric Resch.

Stone Creek plans to continue auditing different parts of the customer journey and using performance data to decide what enters the testing roadmap.

“We let the data guide our priorities. We regularly audit the site (or pages) to see which areas are underperforming. We then form our hypothesis and build the new variant, and set up the A/B test. This is our monthly cadence and it works extremely well.”