PartsTree rebuilds a legacy ecommerce ecosystem with Shopware to support massive scale and complex product data

Rebuilding Commerce for a Highly Specialized Business

A division of Hill Country Outdoor Power, PartsTree operates at the intersection of ecommerce and technical product discovery. As a supplier of replacement parts for outdoor power equipment, the company serves both individual consumers and professional repair businesses with a catalog that spans millions of components across hundreds of brands.

What sets PartsTree apart is not just the scale of its catalog, but the way customers interact with it. Rather than browsing traditional product categories, users begin with a specific piece of equipment, navigating through detailed schematics consisting of exploded parts diagrams, images and interactive hot-spots to research, identify and confidently purchase the exact component they need, often down to an individual bolt or washer. This workflow allows customers to move from diagnosis to order in a single, seamless experience.

For more than two decades, this experience was powered by a custom-built system. Over time, however, that system became increasingly difficult to scale, maintain, and adapt.

When a Legacy System Becomes a Constraint

PartsTree's previous platform was not a commercial ecommerce solution. It was a custom-built ecosystem which grew organically over 25 years, designed to support a unique data model and customer experience.

At its core was a loosely structured document database that had continuously evolved to accommodate new brands, products, and edge cases. While this approach provided flexibility early on, the progression over time accumulated massive tech debt, resulting in significant challenges for maintainability and migration. Naturally,

much of the system's logic was convoluted and contained numerous undocumented exception paths, and inconsistencies in product data, naming conventions, and relationships had accumulated across hundreds of manufacturers – exactly what one gets when a small crew of developers are working against the clock for 25 years without meaningful opportunities to refactor and reassess.

As a result, even basic changes became difficult to implement. Scaling the system to support new brands, markets, or business models was increasingly impractical. At the same time, PartsTree's growth trajectory, supported in part by private equity investment, required a more modern, scalable foundation that could seamlessly integrate new brands, expand SKU counts, and address potential roll-up opportunities. The company needed a platform capable of supporting not only its existing complexity, but also its future ambitions.

Transforming Millions of Products and Relationships

The defining challenge of the project was the sheer scale and structure of PartsTree's data. The catalog includes 4.4 million products, 462,000 equipment model categories, and 1.5 million schematics consisting of exploded parts diagram, images, and interactive hot-spots. These diagrams connect individual components to specific models, creating a dense network of relationships. Including all relationship mappings and supporting entities, the system encompasses approximately 105 million relationship records.

Migrating this data required bridging two fundamentally different systems. The legacy environment relied on a loosely structured document database, while Shopware requires a structured MySQL database with significant referential integrity constraints.

The complexity was further compounded by inconsistencies across catalogs from different manufacturers. Each brand defined products differently, using its own conventions for part numbers, attributes, and categorization. Many edge cases and data exceptions surfaced only during ingestion, requiring continuous refinement of the mapping logic.

Importantly, the legacy system remains (for the moment) the active source of truth for PartsTree’s catalog operations. Altus Nova designed and built a real-time data synchronization architecture to keep the two systems aligned: change events from MongoDB are captured via AWS Lambda and staged to DynamoDB, triggering two sync flows — parts and inventory updates every five minutes, heavier catalog changes nightly. The architecture listens specifically for changes rather than re-evaluating the full catalog each cycle, keeping the process efficient at 105M-record scale. Rather than a one-time migration, this became a living integration ensuring the storefront always reflects current data.

Designing a Non-Traditional Ecommerce Experience

A central requirement of the project was preserving and enhancing PartsTree's schematic-centric product discovery experience, which is a necessity when researching required parts for sophisticated engines. Unlike standard ecommerce platforms where products are organized into categories and filtered by attributes, PartsTree's experience begins with a specific piece of equipment.

This approach required a fundamentally different data model. Altus Nova introduced custom entities for diagrams and their relationships to products and equipment models, and built the IPL (Illustrated Parts List) functionality entirely custom — consuming hotspot positioning data from PartsTree’s proprietary diagram utility and layering fully interactive hotspot links over OEM diagrams. Categories were repurposed as organizational structures for equipment rather than traditional merchandising hierarchies. These adaptations, made possible by Shopware’s extensible architecture, support PartsTree’s specialized experience without compromising performance or maintainability.

A Platform Built for Flexibility and Control

Following a structured evaluation process that included multiple vendors, PartsTree selected Shopware based on its flexibility, extensibility, and strong administrative capabilities. Key differentiators included the platform's ability to accommodate custom data models and its rule-based automation tools.

Shopware's Rule Builder and Flow Builder played a critical role in replacing complex custom logic. For example, PartsTree manages a highly detailed shipping matrix, essentially a 300-line spreadsheet of logic, with numerous variables including geography, customer type, delivery constraints, and carrier availability. Instead of hardcoding these rules, the team implemented them within Shopware's rule engine, allowing business users to manage and update logic without development support.

Similarly, order workflows were automated using Flow Builder, enabling dynamic routing, validation, and exception handling based on predefined conditions. This shift significantly reduced reliance on development resources and improved operational agility.

Importantly, the platform selection was made with a three-to-five year roadmap in mind, not Phase 1 requirements alone. Shopware’s native B2B suite was a substantial draw: PartsTree’s ProAdvantage professional customer program requires distinct pricing, ordering, and account management capabilities that Shopware is architected to support natively. Multi-site and multi-store capabilities were equally deliberate — the platform was selected with brand expansion, new channels, and potential white-label deployments already on the horizon, meaning those investments were accounted for in the foundation rather than deferred to a future re-platforming project.

The headless storefront architecture — built on Nuxt/Vue.js, customized from Shopware’s headless starter library — reflects the same logic. As PartsTree adds new brands, CMS integrations, or additional front-end experiences, the headless layer allows those without rebuilding the commerce core. The Vue.js framework also aligns with Shopware’s admin layer, so storefront and admin customizations share a common codebase, reducing future complexity.

Replacing the Order Management System

The most visible outcome of the project is the storefront. The most consequential may be what happened behind it.

PartsTree’s order management system was the oldest and most fragile component of the legacy codebase— the layer where order states, exception handling, and warehouse handoff logic had accumulated over two decades with minimal documentation. Replacing it was not optional, but it also could not be rebuilt as custom software without recreating the same problem a decade later.

Instead, the team rebuilt order management inside the Shopware Admin, using order states, tags, and Flow Builder rules to reconstruct business processes that had previously existed only in application code. Orders are now evaluated against holding and routing conditions before they reach the warehouse rather than after, with exceptions surfaced to customer service through tags rather than kicked back through a manual queue. Business users can adjust these rules directly.

Operating the Shopware admin at PartsTree’s scale required significant joint engineering effort. Administrative tooling designed around conventional catalogs behaves differently against 4.4 million products and the volume of order states and tags PartsTree’s workflows require. Altus Nova and Shopware worked through performance characteristics, configuration patterns, and extension points together, with Shopware’s team contributing code reviews and platform guidance as the requirements pushed into territory the admin had not previously been asked to cover at this scale.

The result is not a bespoke order management system. It is a supported, configurable platform capability that PartsTree’s own team can operate and extend—replacing the single most brittle piece of a 25-year-old codebase with something the business can maintain itself.

Building a Modern, Integrated Architecture

The new solution combines a headless storefront with Shopware's backend, supported by a microbatch and message-centric middleware tier that continuously synchronizes updates from the legacy catalog system. While the legacy system remains the source of truth for product data, the new architecture enables that data to be structured, normalized, enriched, and presented within a modern ecommerce framework.

The migration also preserved continuity for PartsTree's existing customer base. Approximately 700,000 legacy customer accounts were migrated to Shopware, including stored addresses and vaulted Braintree payment methods, so returning customers could sign in and check out without re-entering payment or shipping details. Historical order records remain in the legacy system, with order history on the new platform beginning at launch a deliberate scoping decision that kept the migration focused on the data customers actively use to transact.

Key integrations include a custom Braintree payment implementation, which also enabled new wallet payment methods like Apple Pay, Google Pay, and PayPal that PartsTree could not previously offer. Additional integrations include tax calculation via TaxJar, real-time shipping and address validation through UPS, USPS, and FedEx, and search powered by OpenSearch. Hosting is provided by Webscale, another key partner in the migration process.

"We would come up with 'here's the thing we want to do' and Webscale would help us figure it out. We would work together to divide, define the strategy and build the pieces. There was some debugging and some testing and some code writing, all teams sharing the responsibility." — Sam Bogart, Director of Technology, Altus Nova

A Collaborative Implementation

The project was completed over approximately ten months, following an initial planning phase that included roadmap development, requirements gathering, and system design.

Given the complexity of the data, the uniqueness of the business model and the significantly customized user experience requirements, the implementation required close collaboration between PartsTree, Altus Nova, Shopware, and supporting technology partners. Although Altus Nova was implementing Shopware for the first time, the Shopware team provided proactive code reviews and architectural guidance throughout the engagement.

"Shopware was absolutely committed to making this very successful, taking our requirementsand making changes in the core Shopware codebase, building plugins for us so that we were not having to customize core code, pushing on their extension partners to make updates to the plugins that we needed, being on calls on a weekly basis, and being on a shared Slack channel daily. Having interactions with them was extremely valuable to this whole process." — Chris Hedge, Product Officer, Altus Nova

Teams worked together in real time to address data inconsistencies, infrastructure requirements, and technical challenges. This collaborative approach was essential to navigating the project's complexity and ensuring a successful outcome.

How Altus Nova Managed Migration Risk

Migrating 105 million relationship records from a 25-year-old custom system—without taking a live ecommerce business offline—required systematic validation, exception handling, and phased traffic introduction.

Altus Nova introduced significant validation logic covering every entity and relationship in the system—every part, every model, every hotspot. Pre and Post validation were performed for each synchronization batch, with notifications and opportunity for human intervention for any integrity issues identified: JSON representations of model pages were extracted from both the legacy and Shopware environments and compared at the data level, confirming that what a customer would see was correctly represented in the new catalog. This approach enabled full-coverage validation at a scale manual inspection could never support.

PartsTree’s 25-year catalog contained both known and unknown data quality exceptions. Known exceptions were built into the migration requirements from the outset. During development, Altus discovered several dozen additional exception cases affecting thousands of entities—most were conditions the PartsTree team recognized when raised, while a few represented genuinely new discoveries. The result: the Shopware catalog faithfully reflects the source catalog with models and parts correctly presented through an improved user experience.

The legacy site remained live throughout the engagement. Altus stood up Shopware site as an “early access” parallel environment, enabling side-by-side comparison and a controlled, risk mitigated transition at a pace the PartsTree business and customers both felt comfortable with. The new site launched first as an invite-only experience for established B2B customers; as stability was confirmed, access expanded via opt-in banners. Traffic can be returned to the legacy site at any point—a controlled, reversible transition rather than a hard cutover.

"Migrating 105 million relationship records off a “home grown” system we'd been evolving for 25 years was never going to be simple. What mattered was how Altus Nova worked through it with us — they stayed accountable, kept us informed at every stage, and understood our catalog logic well enough to know which edge cases actually mattered. They built through Shopware's supported extension points instead of forcing custom core changes, so what we ended up with is a platform we can maintain, delivered by a team that stuck with us until it was right.” - Drew Dutton, CTO, PartsTree.com

From Manual Processes to Operational Agility

The platform launched first to established B2B customers, with access expanding to all visitors as stability was confirmed. The operational impact is concrete. On the legacy system, any change—including a text edit—required a developer to implement and a full production deployment to release. Catalog updates ran as batch processes approximately twice daily. The internal development team managed the website alongside catalog management, warehouse tooling, and a legacy ERP, leaving limited bandwidth for ecommerce improvements. Reluctance to release was as much about perceived risk as capacity: releases on the legacy platform had a history of introducing problems.

On the Shopware platform, shipping rules, order holding logic, and CMS content are manageable by business users directly — without a development ticket or a deployment. Catalog updates now sync within five minutes for parts and inventory. Webscale’s hosting infrastructure provides isolated environments for safe testing and rapid iteration, reducing the risk of every release. The result is a platform where the operations team has real control over business logic, and the development team can focus on building new capability rather than maintaining what exists.

"Our platform is built to make most Shopware deployments feel invisible, but PartsTree needed deliberate tuning—and that's where Shopware's model, paired with Webscale, has an advantage over more boxed-in SaaS platforms that don't let you customize infrastructure to this extent. Altus Nova partnered closely with us to represent PartsTree's particular needs, pressure-test options, make infrastructure tradeoffs explicit, and land on the right balance of scale, resiliency, and cost. They ran coordination across teams with real operational discipline, which made it easier for Webscale to dial in to the right fit - vs the standard menu of default templates we have available to Shopware clients." — Jay Smith, CEO, Webscale Networks

A Foundation for Growth and Expansion

With its new platform in place, PartsTree has a roadmap extending well beyond the consumer storefront. The next major workstream is B2B: implementing Shopware’s native B2B capabilities for PartsTree’s ProAdvantage professional customer program, serving repair businesses, dealers, and fleet operators with distinct pricing, ordering, and account management requirements. Because Shopware’s B2B suite was a primary factor in platform selection, this is a development effort on an architecture already designed to support it, not a re-platforming exercise.

Multi-brand and channel expansion is a near-term workstream. New branded stores, segment-specific microsites, and potential white-label deployments can be stood up on the existing Shopware instance, keeping operational overhead low while expanding reach. AI-driven initiatives are also on the roadmap—improving part findability, automating customer support, and scaling content curation across a catalog too large to manage manually. Altus Nova continues as the implementation partner across these initiatives, working in parallel with PartsTree’s internal team as it builds platform capability.

"This is the most performant Shopware site I've ever seen." — Jason Nyhus, President & General Manager, Shopware North America