Sunday Finds 10-15% Conversion Lifts in Weeks

The Challenge

Sunday is Europe's leading B2B merchandise platform, the system companies like Cloudflare, HubSpot and Deel use to manage branded merchandise across HR, marketing and sales. Its buyers are enterprise teams, so the website has to do real selling.

But Sunday's marketing team is lean and its traffic is moderate, the exact profile where traditional A/B testing falls apart. Structured experiments needed developer resources Sunday didn't have, and waiting months for statistical significance would have slowed a fast-growing business. Sunday needed to keep optimising its messaging and sign-up flow continuously, without adding headcount or pulling the product team off the roadmap.

The Solution

Sunday chose Dalton to run AI-powered experimentation on the landing pages that matter most to enterprise buyers. One JavaScript snippet, and Dalton started analysing pages and proposing experiments on:

• Value-proposition messaging tailored to each buyer persona (HR, marketing and sales)
• Page-structure changes to reduce friction in the sign-up flow
• Content matched to traffic source and search intent

Changes deployed automatically, with no duplicate pages and no engineering tickets. Top-performing variants picked up more traffic while weaker ones were phased out, so there was no waiting weeks for significance. And every week, new ideas surfaced from visitor behavior, keeping the site learning on its own.

The Impact

Within the first month, Sunday had found messaging that measurably moved the needle: 40% more product views, a 14% lower bounce rate, and 10% more click-throughs to demo requests.

"We're finally able to test at the pace our business demands. Our website now adapts as quickly as our campaigns do."

The continuous optimization approach transformed how Sunday views its website, from a static asset requiring periodic updates to a dynamic growth engine that improves with every visitor interaction. For a B2B platform selling to enterprise buyers across Europe, being able to test multiple value propositions at once has accelerated its learning cycle significantly.

Why it worked

Sunday's real constraint was capacity. Every test competed for the same scarce developer time, and with moderate traffic a traditional test could run for months before it said anything. Dalton removed both limits at once. It proposes and ships experiments without engineering, and its bandit algorithm reads results continuously instead of waiting for a fixed sample size, so Sunday could learn from modest traffic in days rather than quarters. Instead of rationing a handful of tests a year, Sunday now runs them constantly, and the site adapts as fast as its campaigns do.

"We got meaningful insights in weeks that would have taken us months using traditional A/B testing."