Svibe: CRO & A/B Testing on Shopify
Continuous experimentation program on Shopify for an intimate wellness and sexual wellbeing brand. 13 A/B tests run with Bayesian engine and RPV as decision metric.
+33.9% Revenue per visitor (RPV)
Continuous experimentation program on Shopify for an intimate wellness and sexual wellbeing brand. 13 A/B tests run with Bayesian engine and RPV as decision metric.
Svibe is a intimate wellness and sexual wellbeing brand selling online through Shopify. Boost designed and executed a continuous experimentation program with 13 controlled A/B tests on their main product pages, enrolling real visitors. The decision metric was revenue per visitor (RPV), not simple conversion rate, to ensure every change generated real value.
The challenge
Svibe needed to optimize their product page conversion without losing the brand's premium identity. The challenge was finding which changes drove more revenue per visitor, not just more clicks on the buy button. With two main products and limited traffic, every test had to be rigorous to avoid wasting sample size.
Our solution
We implemented a continuous experimentation program on Mida with a Bayesian engine. Each test was evaluated by RPV (revenue per visitor), not conversion rate. The threshold for declaring a winner was high (90-95% probability) to avoid deploying noise as if it were signal.
Get startedOverall results
Revenue per visitor (RPV)
Conversion rate
Overall store conversion
Strategies implemented
Some of the highest-return strategies executed in this project.
Dark sensual product page redesign
Action
We redesigned the above-the-fold of the flagship product page with a dark, sensory layout aligned with the brand identity. The goal was to improve engagement and perceived value.
Result
The program's winning test: +33.9% RPV with 99.6% Bayesian probability. Validated .
Copy with emotional accelerators
Action
We incorporated emotional copy into the product page, appealing to sensations and benefits beyond the product's technical specifications.
Result
The best variant pointed to +13% RPV, a promising result although it did not reach the significance threshold with the available sample.
Per-product validation: what wins in one doesn't win in another
Action
We replicated the winning concepts from the main product on the second product to verify whether the learning was transferable.
Result
The same dark sensual concept lost twice on the second product. Key lesson: each product is validated separately, never assumed.
Conclusions and learnings
The dark sensual redesign of the main product proved to be the only change that cleared the deployment bar with near-total certainty.
Deciding by RPV instead of conversion rate prevented deploying changes that increased clicks but destroyed value.
What wins on one product cannot be automatically copied to another: always validate separately.
Overall store conversion went significantly during the program period, coinciding with the winner's implementation.
The Svibe program demonstrates that rigorous experimentation, with real business metrics and high thresholds, is the path to generating sustainable impact in ecommerce.
Get started
Want results like these?
Optimize your results and give a boost to your digital business with impactful actions.
Get started