Anyone who works with digital products knows that launching or redesigning a feature involves more than having a good idea. You need to understand how that change affects the user experience, the purchase journey, and, of course, business results.

In many cases, feature changes are shipped directly to all users, but that approach carries risks: an improvement designed to simplify the journey can, for example, create distractions at a critical conversion moment or introduce unexpected friction.

Data-driven product teams avoid this kind of risk by testing hypotheses before a full rollout. AB tests make it possible to validate product decisions in production, with real users, and understand the impact of each change before scaling it.

But for this to work in practice, experimentation needs to be simple, fast, and integrated into the product development workflow.

Printi is one of Brazil's largest online print shops, with a mission to revolutionize the print market through technology, design, and innovation. The company stands out by making high-quality print products more accessible, offering competitive pricing, reliable turnaround times, and flexible solutions that accommodate everything from large orders to one-off or urgent requests. With this positioning, Printi aims to be a go-to partner for businesses of all sizes and needs.

With a complex digital product and a purchase journey that serves varied customer profiles, validating changes before rolling them out to the entire user base is an essential practice.

The challenge: launching a new feature without hurting conversion

The Printi team identified a clear opportunity to improve the purchase journey: making it obvious to users that they could download quotes directly on the website.

This feature is especially useful for companies that typically buy in larger quantities and need to submit quotes for internal approval before completing an order. Until then, because the feature wasn't easy to find, many of these customers were reaching out to support channels to request that kind of document.

The change had two goals: simplify the purchase journey by giving users more autonomy, and reduce the volume of support tickets related to quote requests. There was, however, an important concern. The feature could reinforce a drop-off point in the purchase flow, especially at critical moments in the journey.

Before rolling it out, the team needed to answer a few questions:

• Would the feature change affect conversion rates on the product page or in the cart?
• In what position or format should it appear to add value without getting in the way of the purchase?
• Would it actually help take the pressure off the support team?

To answer these questions, Printi decided to run a test with multiple variations.

Experimentation applied to product

Using Croct, the team set up experiments to test the feature in two strategic areas of the site, separately: the PDP (product detail page) and the cart.

On the PDP, the feature was being introduced for the first time, so the experiment included four variants, making it possible to evaluate not just whether the feature should be there, but also how it should be presented in the interface.

This approach allowed the product team to test multiple hypotheses at the same time, such as whether this new option would work better as a secondary action or as a visual highlight, and whether user behavior would shift depending on the page context.

By validating these variations in production with real users, the team was able to observe how each approach influenced behavior within the page, without having to assume upfront which solution would be best.

During the four PDP variant tests, the critical point was the change to the first fold. Our assumption was that the most aggressive variant would hurt conversion, but the result surprised us. That version ended up having the most neutral impact, which invalidated our initial belief that the fold's design would dictate customer behavior in the funnel.

Learning before the rollout

This kind of approach brings an important benefit to product teams: it reduces the risk of shipping feature changes that don't perform as expected.

Rather than assuming a change will necessarily improve the experience, Printi treated it as a hypothesis to be validated.

By running the experiment, the team was able to:

• Assess whether the feature interfered with conversion on critical pages in the journey.
• Understand how different ways of presenting the feature influenced user interaction.
• Identify which variant struck the best balance between usefulness and impact on the purchase flow.

These insights allow rollouts to happen with much greater confidence, since decisions are grounded in real user behavior. Instead of shipping the change to all users and then trying to fix any side effects after the fact, Printi was able to learn from real data and refine the solution while still in the experimentation phase.

Beyond guiding product decisions, this kind of approach also makes it easier to align different teams across the company. When decisions are backed by experiments and data, conversations stop relying solely on internal assumptions and start being driven by evidence of actual user behavior.

Wrapping up

The Printi case shows that AB testing isn't just relevant for marketing and growth. It's also a powerful tool for product teams that want to optimize features with greater confidence.

By testing the quote download feature before the full rollout, Printi was able to validate hypotheses, reduce risk, and better understand how users interact with the product. This kind of approach turns experimentation into a natural part of the development process. Instead of shipping changes and hoping for the best, the team learns from users and makes more informed decisions.

In the end, testing before launching not only improves the user experience, but also helps the product evolve in a more consistent, data-driven way.

If you also want to make product decisions backed by data, create your free account and start testing with Croct today.