From idea to live before lunch: how CVC launched an AB test from scratch in 20 minutes
When experimenting stops being a project and becomes part of the workflow, what changes isn't just the timeline, it's what the team feels empowered to test.
Every team running an online sales channel lives with the same routine. Change requests come from all directions: a new banner, a different copy, one more highlight on the page. In isolation, each request seems small. Added together, they become the difference between an operation that learns and one that guesses.
The question that separates data-driven companies from the rest isn't whether it's worth testing. It's how to test without missing the business window. Every idea that takes weeks to go live pays a toll in two currencies: team time and opportunity cost.
That was exactly the knot CVC set out to untie.
CVC is the largest travel operator in Latin America, with over five decades in the business, a presence in more than 1,000 physical stores, and a robust digital operation. At that scale, every interface detail becomes a lever: what looks like a small change on a page, multiplied by daily traffic, turns into revenue lost or won.
Test individual elements or full journeys with advanced AI-powered segmentation, great performance, no-code rollouts, and flicker-free experiments.
The challenge
The CVC team wanted to validate a seemingly simple hypothesis: test a label variation on the CTAs that direct users toward the in-store purchase flow. A small surface-level adjustment, with potentially meaningful impact on the conversion flow.
The hard part wasn't the change itself. It was everything that normally comes with it:
- Aligning the hypothesis across the teams involved
- Handling the initial technical integration
- Configuring the experiment
- Shipping the test to production safely
In most companies, that path runs through briefings, tickets, a development queue, and release cycles. The side effect is well known: the test is dead on arrival because the window in which the hypothesis made sense has already closed.
The cost of a slow test isn't the time spent. It's the number of tests that never happen because the team learned it "wasn't worth it."
How CVC shortened the path
With Croct, the entire flow was condensed into a single morning. The team ran the full sequence without opening a single dev ticket, joining a release queue, or handing off the test operation to another team.
On the same morning, the experiment team itself:
- Defined the hypothesis and quickly aligned with the teams involved
- Handled the initial technical integration
- Configured the experiment directly in the platform
- Published to production with the test running live
From the first click to the variation live for real users: about 20 minutes. Before lunch, the test was already active and collecting data.
Before and after, side by side
| Traditional flow | Workflow with Croct |
|---|---|
| Cross-team briefing, technical validation, dev ticket, prioritization queue. | Hypothesis aligned the same morning, no dev queue dependency. |
| Technical integration tied to the product release cycle. | Initial integration handled by the experiment team itself, in minutes. |
| Experiment setup dependent on manual implementation. | Test configured directly in the platform, no new code. |
| Published to production days or weeks after the hypothesis emerged. | Test live before lunch, on the same day. |
What this unlocks in practice
Reducing the time between idea and production sounds like an efficiency improvement. It's more than that: it changes the type of hypothesis the team feels comfortable raising.
1. The cost of testing a "small" idea drops
When shipping an experiment costs a full sprint cycle, teams only test what feels worth the effort. Smaller ideas, potentially good ones, die in the queue. With a setup that takes minutes, those hypotheses come back to the table.
2. The team regains autonomy without losing control
Autonomy without governance turns into chaos. The flow CVC uses keeps the product team in charge of the operation (alignment, configuration, and publishing) while maintaining visibility and control over what's live.
3. Experimenting stops being a project and becomes a routine
This is the deeper effect. When the cycle is measured in minutes rather than weeks, experimentation stops being a one-off initiative with its own ceremony and becomes how the team works. It's the difference between "running a test" and "being a team that tests."
Wrapping up
The story of this case isn't about the outcome of the label variation. It's an operational realization: at CVC, an idea that came up in the morning was live before lunch.
In a market like travel, where the opportunity window shifts with every season, holiday, and competitor move, that speed isn't a luxury, it's how the operation responds to reality. Shortening the time between hypothesis and data ultimately shortens the time between guessing and learning.
And learning that arrives fast is what separates those who decide based on what they know from those who decide based on what they assume.
If you also want to make product decisions backed by data, create your free account and start testing with Croct today.