The Challenge
When Eppi, a family-owned jeweller from Slovakia, first launched in 2011, their mission was clear: create beautiful, personal, and sustainable jewellery. Over the years, that vision helped them expand into the Czech Republic and Germany, offering everything from engagement rings to bracelets – all made from recycled materials and lab-grown diamonds.
But growth brought complexity. Eppi’s marketing spanned a wide range of channels – Meta, Google Ads, Pinterest, Microsoft Ads, YouTube, LinkedIn, Snapchat, and more. Their customers’ journeys were rarely quick; choosing the right engagement ring or investing in a unique gemstone bracelet could involve weeks of research, multiple visits, and around 10 different touchpoints across platforms and devices.
Yet, their measurement framework hadn’t kept pace. Eppi relied on a last-touch attribution model, which rewarded only the final interaction before purchase. This approach ignored much of the customer journey – impressions, earlier clicks, and multi-device behaviour – leaving major blind spots in performance reporting. As CEO Matej Sukovský explained:
“Our customers take their time. Their journey isn’t a single click; it’s a sequence of touchpoints – often nine or more. Understanding the contribution of each one is critical, but far from simple.” – Matej Sukovský, CEO Eppi
The outcome? A significant portion of ad spend was being invested inefficiently, with budget decisions driven by incomplete data.
The Solution
In late 2024, Eppi partnered with Roivenue to rebuild their measurement from the ground up. The first step was implementing Roivenue Measurement to track traffic sources, user behaviour, and conversions. This was combined with cost data from advertising platforms and order management system data – including margins, cancellations, and returns – to establish Roivenue as Eppi’s single source of truth. With this holistic data foundation, Roivenue could move beyond basic channel attribution and evaluate the true role of each touchpoint in driving conversions.
To close the biggest measurement gap – multi-device journeys – Roivenue introduced its Path Stitching methodology. This probabilistic AI-based feature connected separate sessions from different devices, enabling a full view of the journey even when cookies differed. The impact was dramatic: the average number of touchpoints per journey rose by up to 80% once previously invisible steps were revealed.
The Impact
With the complete data pipeline in place, Roivenue deployed its AI-driven attribution model powered by recurrent neural networks (RNN). This model evaluates the entire customer journey at a granular level – right down to individual campaigns, ad groups, and creatives – enabling smarter, data-driven decisions across the funnel.
One of the most striking insights emerged when the model revealed just how much traditional attribution had been undervaluing certain channels – especially Facebook. Under last-touch attribution, Facebook’s performance was heavily underestimated. In reality, it delivered 505% more impact than last touch reported, since impressions and earlier clicks were ignored. Roivenue’s synthetic impression attribution methodology captured these hidden contributions, transforming how Eppi evaluated and invested in the platform.
Another case came when comparing Direct performance under last-touch attribution versus Roivenue’s AI model. Because the AI model accounts for all preceding touchpoints (including those on different devices), it reduced Direct’s credited performance by 66%. This revealed which channels were truly driving revenue, enabling better budget allocation.
The Growth
The results spoke for themselves: comparing the first half of the year before Roivenue (H1 Year 0) to the first half of the first year with Roivenue (H1 Year 1), Eppi achieved 124 % increase in mROMI.
124% mROMI (Margin Return on Marketing Investment)
Another area where growth was clearly visible was conversions, which increased by 25% year-over-year. Naturally, this was reflected in net revenue as well, which rose by 21%.
+25% conversions
+21% net revenue
These gains reflected not just improved measurement, but Eppi’s willingness to act on the insights – reallocating budgets toward the campaigns, creatives, and channels that the AI model identified as true drivers of growth.
The Conclusion
Eppi’s transformation shows what’s possible when a brand moves beyond outdated attribution models and embraces a full-funnel, multi-device view of the customer journey. By combining a robust measurement framework with advanced AI modeling, Roivenue helped Eppi see and act on the real levers of growth. For brands with complex journeys and high-value purchases, the lesson is clear: if you can’t see every step your customer takes, you can’t invest with confidence. Eppi can – and the results prove it.
This success story is just one example of what’s possible. Discover more client case studies to see how we’ve helped other businesses reach their goals.
Take your marketing to the next level!
Schedule a call with our team today.
