How Life is Good Built a Culture of Experimentation and Let the Data Lead

Life is Good spent most of its history as a wholesale and retail brand: 52 of its own stores, plus shelf space at many popular retail partners. The e-commerce business runs on a print-on-demand model with a fleet of direct-to-garment printers in their  New Hampshire factory. That physical reality puts margin under constant watch.

During the pandemic, e-commerce went from a minor share to about two-thirds of the company's total revenue. Fran Middleton arrived to oversee the business in October 2025, inheriting a team of 14 people inside a roughly 250-person company. He believed a small team with the right tools and the right data could run circles around a much larger one to scale the business even further. Fran saw an opportunity to bring more rigor to how the team evaluated performance. "We wanted to be able to measure and quantify what was working and what wasn't, both the wins and the lessons." One of the first things Fran put in the 2026 budget was an A/B testing tool. The team evaluated options, picked Shoplift, integrated it in early January, and started testing.

Building a Testing Habit From Scratch

Fran came in focused on building a culture of experimentation and started by focusing on low-lift optimizations that needed no developer time. Early on, the goal wasn't any single result. It was more about getting the team to trust the method and build the testing muscle.

That happened faster than Fran expected. The early tests produced quick wins, and the backlog of ideas grew as people across merchandising, marketing, and web ops started adding their own. A short starter list turned into a queue everyone wanted into. "The whole team is all in," Fran said.

The setup stayed lightweight. Marketing owns test strategy and hypotheses alongside Fran, and execution pulls in whoever the test needs: web ops for anything that ships through Shopify configuration, engineering for heavier work, design and UX for the rest. The backlog is scored by expected upside against effort, which is how the team decides what to run next.

Questioning the Defaults

Once the testing muscle was in place, the team pointed it at a bigger target: the assumptions baked into how the business had always run. Decisions that once rested on gut feel now get framed as hypotheses. Some defaults survive the scrutiny. Others turn out to be leaving real revenue on the table.

Two examples show the range.

Proof Point One: Rethinking the Shipping Threshold

Shipping policy was a settled question at Life is Good. An earlier threshold had coincided with a sharp drop in sales, and free shipping had been the default ever since. What reopened the conversation was outside the company's control. As 2026 began, a USPS fuel surcharge tied to the conflict with Iran started cutting into margin, and on a print-on-demand model where every unit already carries cost, the economics of shipping deserved a fresh look.

Life is Good first moved to a $25 threshold and watched conversion rate dip, but average order value rose enough to keep the math working. With Shoplift in place, the team didn't have to guess on the next move. They ran a controlled test: raise the threshold from $25 to $35 and measure the trade-off.

The test hit statistical significance in 13 days and called a clear winner. Conversion rate dipped, as expected, but average order value rose 6.3%. Revenue per visitor, the number that actually nets the trade-off, rose 1.3% against a conversion-rate change of -4.7%. The gain on the metric that mattered exceeded the friction the higher threshold added.

The data revealed two customer behaviors at once. Many shoppers added a second item to clear $35, which pushed average order value up. And a meaningful share of single-item shoppers absorbed the shipping cost without abandoning their purchase. This was a strong signal of how much the product itself resonates.

The test also surfaced an upside for customers. When the team runs a clearance event or a seasonal promotion, adding free shipping creates a real reward rather than blending into an always-on default. And the incremental shipping revenue more than offset the carrier surcharge that prompted the question in the first place. A change made to protect margin paid for itself in a way no one had forecast.

Proof Point Two: Making the Fit Difference Clear

The same instinct to turn an assumption into a hypothesis pointed the team at what customers were telling them after the sale.

Over the past few months, customer feedback surfaced a theme around size and fit. Some shoppers noted that a shirt ran shorter or wider than they expected. On the surface it read like a product question. The team looked closer and formed a different hypothesis.

Life is Good sells several tee silhouettes: a boxy cut, a shrunken cut, a regular crew, and more. The team's read on the feedback was that some shoppers thought they were buying a standard crew when they had actually ordered a boxy or a shrunken fit. The gap wasn't the garment, it was the expectation. The website had room to make the difference between silhouettes clearer at the moment of choice.

So they built a test around it. The new variant places each tee silhouette on a model at the top of the product listing page, so the distinct fit of every cut is obvious before a shopper commits. The hypothesis is simple: show the difference clearly, and more orders land the way customers expect.

The team is still observing the trend rather than calling a result, however early indications are positive. When an order matches what a shopper pictured, they're happier with it and less likely to send it back.

It's the same test-and-learn logic applied to a signal most brands would file under product feedback: take the assumption, frame it as a hypothesis, and let the page prove or disprove it.

Testing as the Default

While individual test results dominate the headlines in Life is Good's e-commerce meetings, the underlying shift is cultural: asking "what if we tried it differently?" has become the default. The payoff has shown up in three places:

• An e-commerce team that now runs on data, with ideas coming from everyone on the team
• Revenue growth from revisiting old assumptions
• A lean team using data to make faster, more confident decisions as the business continues to scale

As Fran describes it, Shoplift is a "simple, easy-to-use platform that allows you to A/B test ideas across all parts of the funnel." And by their own measure, it has "more than paid for itself in terms of the tests we've run and the value it's created for our business."

For any brand sitting on a practice that has gone unquestioned in years, Life is Good's approach is worth copying: test it, watch the metric that actually matters, and let the result decide what stays.

‍