CLIENT CASE Aveve
From Proof of Concept to omnichannel activation: how Aveve redesigned its customer engagement stack.
About
Aveve
Aveve is Belgium’s retail chain for garden, agriculture and pet products, part of the Arvesta Group. Its stores serve a broad customer base (pet owners, hobby gardeners, agricultural professionals) with a loyalty programme counting more than one million active cardholders.
The business is strongly seasonal. A customer who buys cat food in January may buy a lawn mower in April. That breadth of purchasing behaviour across categories is exactly what makes personalisation both difficult and valuable.
Introduction
Many organisations sit on large volumes of customer data but struggle to translate that into meaningful marketing impact. Aveve was no exception. With over one million loyalty cardholders, the data was there, but the ability to act on it in a structured, scalable way was limited.
A customer who had just bought their first cat litter box received the same message as a decade-long pet owner. Someone actively browsing lawn mowers online received no follow-up. A first purchase in an unfamiliar category, the clearest possible signal of intent, triggered nothing.
So the core question was not whether the data had value, but how to unlock it. What use cases actually drive incremental revenue? What capabilities are missing in the current setup? And how do you move towards a model where marketing is driven by customer behaviour rather than campaign planning?
This case outlines how Aveve moved from a proof of concept to an implementation of a new customer engagement and loyalty platform, and the role Stitchd played in making that transition viable.
"Moving fluidly between strategic thinking and concrete delivery, between the IT side and the commercial marketing side. Being able to work along both of those axes is, for me, one of Stitchd's core strenghts.
Peter De Mey
Marketing Director, Aveve
Project approach
Phase 1: A Proof of Concept was set-up in a single category
Instead of starting with platform selection, the approach began with a pilot project. The scope was limited to the pet food category, which provided a contained environment to test the impact of behavioural segmentation.
Using Azure as the analytical foundation, customer data was consolidated and analysed to identify actionable segments. These included signals such as potential dog ownership, churned customers, and cross-sell opportunities within the category. These insights were then translated into targeted marketing flows.
The objective was not scalability at this stage, but validation. By keeping the scope focused, it became possible to isolate the impact of better segmentation and demonstrate its commercial value without introducing unnecessary complexity.
Phase 2: from validation to organisational commitment
The results of the proof of concept were clear. Personalised campaigns driven by these segments resulted in a sales uplift of 3% - 4.5%.
More importantly, the exercise changed the internal conversation. What had previously been seen as a complex data challenge became a tangible opportunity with measurable impact. This shift enabled a concrete decision at executive level to invest further in a structural solution.
At this point, the focus moved from isolated use cases to building the capabilities required to scale this approach across the organisation.
Phase 3: defining the target architecture
With business validation in place, the next step was to design a future-proof architecture. The goal was to move away from fragmented tooling towards a coherent customer data infrastructure that could support both marketing activation and loyalty use cases.
The architecture was defined as a combination of interconnected components, including a customer engagement platform, a loyalty platform, and a consent management layer. These needed to integrate with the existing data platform, digital touchpoints, and the cash register system.
This design addressed several structural challenges, such as overlapping capabilities, inconsistent data definitions, and limited integration between online and offline channels. It also ensured that privacy and consent management were embedded in the setup rather than treated as an afterthought.
Phase 4: vendor selection based on capability fit
Vendor selection followed the architectural design rather than preceding it. This ensured that the evaluation was grounded in clearly defined requirements instead of feature comparisons in isolation.
The selection process focused on how well potential solutions could support loyalty use cases, integrate with the existing ecosystem, and align with the defined architecture. Total cost of ownership was considered alongside functional fit.
This led to the selection of Voyado, which combined customer engagement, loyalty management, and consent handling into a single integrated solution. This reduced fragmentation and ensured that key capabilities could be managed within one environment while still fitting into the broader architecture.
Phase 5: implementation and data migration
The implementation phase was significantly more complex than the initial proof of concept. Where the PoC had been intentionally small and controlled, the implementation required integration across multiple systems and teams.
Voyado needed to be connected to the cash register, existing data platforms, and digital channels. At the same time, a large volume of historical customer data had to be prepared for migration.
Stitchd’s role in this phase focused on data quality and migration. Before any data could be transferred, substantial cleanup was required. This included resolving duplicate profiles, correcting invalid contact data, aligning identifiers across systems, and addressing inconsistencies between loyalty and transactional data.
Once the data was cleaned, it was mapped to the new data model and migrated into the new environment. This process required careful validation to ensure that customer profiles, loyalty status, and historical interactions were preserved correctly.
In practice, this “move-in” phase determined how quickly the new platform could deliver value. Without proper data quality and migration, the new setup would have inherited the same limitations as the old one.
What changed: enabling behaviour-driven marketing
Following the implementation, Aveve moved from a campaign-driven approach to a more behaviour-driven model. Customer interactions, such as purchases, browsing behaviour, and in-store activity, became direct inputs for marketing logic.
This made it possible to trigger communication based on actual customer behaviour rather than predefined schedules. A first purchase in a new category could initiate a tailored onboarding flow, while inactivity could trigger reactivation campaigns. At the same time, the integration of loyalty data ensured that incentives and rewards were aligned with customer value.
The result was not just more personalised communication, but a more consistent and connected customer experience across channels.
Conclusion: scaling impact requires more than a successful PoC
The proof of concept played a critical role in demonstrating value, but it was only a small part of the overall effort. The majority of the work and complexity sat in the phases that followed, particularly in architecture design, vendor selection, and implementation.
What made this approach effective was the sequence. By validating impact first, Aveve avoided investing in technology without a clear business case. By defining architecture before selecting vendors, the organisation ensured that tools supported the strategy rather than shaping it. And by investing in data quality and migration, the foundation was put in place for long-term success.
For organisations facing similar challenges, the key takeaway is that becoming data-driven is not a single step, but a structured process where each phase builds on the previous one.
CLIENT CASES
More successful cases
