The last click is the most expensive place to lose a customer.

The client

Dry Farm Wines is a subscription wine company focused on healthy, sulfate-free wines. The brand was strong, traffic was healthy, and the product had devoted subscribers. The company’s CEO, Todd, came to us with a specific, well-formed problem: the checkout page was converting below what the rest of the funnel deserved.

The challenge

Checkout is where hesitation is most costly. A visitor who has chosen wines, entered a cart, and started paying has done nearly all the work — and every point of friction or doubt at that moment is a completed sale lost. The task was not to redesign the brand’s checkout on taste, but to find, by experiment, the version that most people actually finish.

The approach

After auditing the full checkout experience, we proposed two areas of experimentation: the cart page and the checkout page itself. Test candidates came from the studio’s documented pattern library, focused on core persuasion elements that research says matter at the point of payment:

• Authority — signals that credible sources stand behind the product.
• Social proof — evidence that other buyers have made, and enjoyed, the same decision.
• Endowed progress — showing buyers how far they’ve already come, so abandoning feels like losing ground.

Each variant ran against the control in structured rounds, and no round was called until it reached statistical significance.

What moved

When the split tests reached significance, the winning combination had produced a 92% increase in completed purchases on the checkout page versus the control. Same traffic, same product, same price — different decisions at the moment of payment.

What this case shows

The highest-leverage testing ground in most funnels is the step closest to money. Small, research-backed changes at checkout compound across every visitor who reaches it — which is why this is often where we look first.