In an increasingly competitive SaaS market, every detail matters. A copy tweak, a positioning shift, or a new hero promise may seem minor, but the impact can be significant.

That's exactly what Cobli discovered when testing a new message on their website. Like any data-driven company, they used AB testing to validate a positioning hypothesis and avoided greater losses by quickly identifying a negative result. For the growth team, turning experiments into practical learnings is essential to refining their conversion strategy.

Cobli is an intelligence platform for fleet management, built for companies that keep the world moving. Founded in 2017 by Parker Treacy and Rodrigo Mourad, the company grows over 100% per year, has a team of more than 400 people, and has raised over R$350 million in investment, led by SoftBank, IFC, and Fifth Wall.

Operating across every state in Brazil, Cobli serves thousands of clients in sectors such as telecom, transportation, energy, e-commerce, and construction. Their solutions combine IoT, artificial intelligence, big data, and video telematics to reduce costs, boost productivity and safety, improve driver behavior, and ensure operational compliance.

Cobli's challenge

Cobli's goal was clear: increase qualified lead conversion on their website.

As the product and brand positioning evolved, the CRO team realized that the main hero copy no longer reflected the value the platform delivered. The hypothesis seemed straightforward and quite common in SaaS: a new, more aligned message could make the value proposition clearer and, as a result, drive more conversions. After all, if the product had evolved, it only made sense for the communication to evolve with it.

The AB test and the surprise

The CRO team ran an AB test comparing the original copy against the new version. The experiment focused exclusively on the text, keeping all other page elements constant to isolate the message's impact.

The result, however, was the opposite of what they expected. The new hero version showed a 30% drop in SALs (sales accepted leads) and 50% fewer demo bookings. The new copy didn't just underperform; it directly impacted critical funnel metrics.

Fortunately, the test made it possible to catch the problem early. Cobli was able to course-correct in time and revert the site to the original version for that audience, preventing an even larger hit on results.

The first learning

In SaaS, every detail matters, and this experiment reinforced a key point for lead generation: copy is never "just copy." It carries expectations, shapes perceptions, and directly influences user decisions. Small changes can have large effects on how users perceive your product, for better or worse.

The first takeaway from this example is that any change needs to be tested. Even adjustments that seem safe and well-aligned with the product's positioning need validation. Intuition and best practices help shape hypotheses, but only data reveals the real impact.

Croct gives us the confidence and autonomy to implement changes on our website that, without an AB testing tool, we wouldn't dare push to production.

The second learning

When the team analyzed the data more closely, they found that the first test's results hid important nuances once broken down by segment. Users coming from paid media and organic users showed distinct behaviors, expectations, and levels of awareness.

This led the CRO team to develop a new hypothesis: would the best path forward be to test different messages based on acquisition channel?

A new experiment

With this new hypothesis, Cobli kicked off a second phase of experimentation. The new test segments users by traffic source:

• Users coming from paid campaigns
• Users coming from organic traffic

Each group now sees a different version of the copy, competing against the original, better accounting for the context and intent of each audience.

What we can learn from Cobli

It's important to understand that AB testing isn't a one-shot deal, and it doesn't exist just to confirm wins. Every experiment generates new insights, raises new questions, and opens the door to more refined hypotheses. The goal isn't just to prove a hypothesis right, but to continuously improve the site's messaging and experience.

Cobli's experience shows in practice how a continuous experimentation cycle is essential for making safer, more informed business decisions.

The ongoing learning from our tests has been key to deepening our understanding of user behavior.

Beyond chasing immediate gains, testing allows you to prevent losses, save development time, avoid rework, and build real knowledge about how your users behave.

If you also want to turn your website into a continuous learning machine, create your free account and start testing with Croct today.