Case Study: Rebuilding a Parts Sourcing and Installation Marketplace

# Case Study: Rebuilding a Parts Sourcing and Installation Marketplace for a Regional Repair Network ## Overview A regional automotive service network operating across five states wanted to modernize how it sourced parts and managed installation jobs. The business supported 120+ partner garages and independent mechanics, handled thousands of SKUs, and coordinated a rotating list of salvage yards and distributors. The old process relied on vendor spreadsheets, manual phone calls, and inconsistent pricing rules. Customer experience suffered, partners struggled to keep inventory current, and leadership lacked trustworthy operational metrics. Webskyne was engaged to design and deliver a unified marketplace-style platform that would streamline the parts lifecycle end-to-end: catalog ingestion, compatibility checks, quoting, ordering, job dispatch, and settlement. We rebuilt the workflow with a web admin portal, partner dashboards, and mobile apps for field technicians. The result was a scalable, data-driven system that cut order cycle time, increased conversion, and introduced reliable performance visibility. ## The Challenge The organization’s ecosystem had grown faster than its tooling. A few problems kept recurring: 1. **Fragmented inventory data**: Each supplier maintained its own format with varying completeness. Some used CSV exports, some used PDFs, and some shared a Google Sheet. Part descriptions were inconsistent and often missing compatibility details. 2. **Slow quoting and approval**: Staff manually called suppliers, verified availability, and created a quote in a legacy CRM. Average quote turnaround exceeded 24 hours and led to lost sales. 3. **Poor installation coordination**: Installing a part required juggling schedules between a garage, a field technician, and the end customer. There was no real-time tracking or proof of completion. 4. **Unreliable metrics**: Leaders didn’t trust reported order volume, job completion rates, or vendor performance. Operations depended on anecdotal updates instead of data. 5. **Security and compliance risks**: Sensitive customer details were being sent over email and shared on spreadsheets without audit trails. The organization wanted to act like a modern marketplace while still preserving the relationships and flexibility that its regional partners depended on. ## Goals We translated the business vision into measurable goals that would define success. The project needed to deliver: - **Faster quotes**: Reduce the quote response time from >24 hours to <2 hours for common parts. - **Higher conversion**: Increase quote-to-order conversion by 15–20% within the first quarter after launch. - **Reliable compatibility**: Provide accurate part-vehicle matching with clear confidence scoring. - **Streamlined installs**: Create a transparent workflow for scheduling and completing installation jobs with photographic proof and customer approval. - **Operational visibility**: Deliver dashboards that show order volume, vendor fill rates, job completion times, and customer satisfaction scores in near real time. - **Scalability**: Support growth from 120 to 250 partner garages without degrading performance or requiring additional administrative headcount. ## Approach We used a phased, risk-reducing approach that combined deep discovery with incremental delivery. The team aligned on five guiding principles: 1. **Standardize data before automating**: A marketplace is only as strong as its catalog. We needed canonical SKUs, normalized metadata, and compatibility logic before real automation could work. 2. **Design for partner adoption**: Vendors and garages were the engine of the network. The platform needed to be simple enough to use with minimal training. 3. **Mobile-first for field techs**: Installation jobs happen on-site. We prioritized mobile workflows for dispatch, geo-verification, and photo capture. 4. **Build with auditability**: Every quote, order, and job needed a traceable history to support disputes and compliance. 5. **Deliver value every sprint**: Instead of a “big bang” launch, we planned for multiple releases that improved specific workflows while the legacy system stayed online. We started with discovery workshops and process mapping, then validated assumptions with real partner data. The architecture was designed for modularity to allow the inventory pipeline, compatibility engine, and dispatch system to evolve independently. ## Implementation ### Discovery & Data Audit The initial phase focused on mapping the current end-to-end flow and diagnosing failure points. We conducted interviews with operators, partner garages, and field technicians. Then we audited 50,000 historical orders to understand quote timing, cancellation reasons, vendor performance, and part categories that drove profitability. The data audit revealed that 37% of orders required manual correction due to incomplete part descriptions or mismatched vehicle compatibility. This was the single biggest contributor to delays. We used that insight to prioritize the data normalization layer and compatibility engine as foundational work. ### Architecture & Core Stack The solution was built as a modular platform: - **Frontend web portals** for administrators and partner garages, built in Next.js for speed and SEO-friendly documentation pages. - **Mobile apps** in Flutter for field technicians and customer-facing job tracking. - **Backend services** in NestJS with PostgreSQL for transactional data and Redis for queue processing. - **Storage and media** using S3-compatible object storage for images and documentation. - **Geolocation** with Google Maps for routing, geo-fencing, and arrival verification. We separated the system into distinct domains (catalog, quoting, orders, dispatch, and reporting) with well-defined APIs. This allowed iterative delivery while minimizing cross-team dependencies. ### Inventory Ingestion & Normalization We created a flexible ingestion pipeline that could accept CSV, XLSX, API feeds, or manual uploads. Each incoming feed was mapped to a canonical schema, then enriched with standardized metadata (part category, manufacturer, condition, and a normalized VIN compatibility mapping). A human-in-the-loop validation queue allowed operations staff to fix outliers without blocking the entire import. The system highlighted missing attributes and required vendors to provide a minimum data threshold before listings could go live. Over time, this improved vendor data hygiene without forcing immediate compliance. ### Compatibility Engine Compatibility was a critical differentiator. We introduced a rules-based engine backed by a vehicle fitment database and a confidence scoring model. When a part matched a vehicle, the UI surfaced a confidence rating and suggested alternatives if the score was low. Operators could override results with a reason code, which fed into a continuous improvement loop. We deliberately made the engine explainable. Every compatibility match showed which attributes drove the result. This reduced disputes and built trust with partner garages. ### Quoting & Checkout The new quoting workflow replaced manual phone calls. Operators could select a vehicle, choose a part, view pricing from multiple vendors, and generate a quote in minutes. The quote included part details, expected delivery time, warranty info, and installation options. Customers received a branded link to approve or reject the quote. If approved, payment and scheduling happened within the same flow. This removed an entire layer of back-and-forth communication. ### Dispatch & Field Technician Workflow For installation, we launched a field technician app that handled: - Job assignment and acceptance - Geo-fenced job start (tech must be within a defined radius) - Photo capture of “before” and “after” states - Customer signature on completion - Automatic status updates to the admin dashboard We also built a dispatcher view that allowed coordinators to balance workloads and track on-time arrival rates. This replaced manual whiteboard scheduling and phone calls.  ### Observability & Reporting We built a reporting layer using a denormalized analytics schema. Key dashboards displayed: - Quote turnaround time - Conversion rates by vendor and region - Installation completion times - Technician on-time arrival - Dispute and refund rates The dashboards were built into the admin portal so leadership could see trends without waiting for weekly reports. We also exposed a lightweight CSV export for deeper analysis. ### Security & Compliance Because the system handled personal information and payment data, we implemented: - Role-based access control for each portal - Audit logs for quotes, order changes, and job updates - Secure tokenized payment processing through a PCI-compliant provider - Automated retention and deletion policies for customer documents This gave the organization confidence that the platform could scale without introducing compliance risk. ## Results The platform launched in phases over 16 weeks. Each phase replaced a specific manual workflow while keeping legacy tools available as a fallback. This reduced disruption to partners while enabling real-time learning. The first visible impact was speed. Average quote turnaround fell dramatically, and partners reported higher customer satisfaction because they could respond on the same day. The installation workflow produced reliable status updates, which reduced inbound support calls. Vendors also appreciated the standardized listing process because it gave them visibility into how their parts were performing. By the end of the first quarter after launch, the organization had enough data to take proactive action on vendor performance and technician utilization. Leadership no longer had to rely on anecdotal updates; metrics were now trusted and widely referenced in weekly operations meetings. ## Metrics - **Quote turnaround time**: Reduced from 26 hours to 1.8 hours for common parts (93% improvement). - **Quote-to-order conversion**: Increased by 18% within 90 days. - **Vendor fill rate**: Improved from 72% to 86% due to better data requirements and availability tracking. - **Installation completion time**: Reduced by 24% through optimized dispatch and geo-verification. - **Customer satisfaction (CSAT)**: Improved from 3.6 to 4.4 on a 5-point scale. - **Operational visibility**: Weekly management reporting time dropped from 10 hours to 1.5 hours. These outcomes validated the decision to focus on data normalization and workflow automation before scaling marketing or expansion efforts. ## Lessons Learned 1. **Data quality is the fastest lever**: The biggest operational gains came from enforcing minimum catalog standards. Even a modest improvement in inventory metadata removed dozens of manual steps. 2. **Explainability builds adoption**: The compatibility engine gained trust because it showed its reasoning. Partners were far more willing to use it when they understood why a match was recommended. 3. **Field workflows need redundancy**: Technicians operate in unpredictable environments. Offline support and automatic retry logic were essential to avoid stalled jobs. 4. **Phased rollouts reduce risk**: Keeping legacy tools available while new features went live prevented resistance and ensured business continuity. 5. **Metrics drive accountability**: Once vendors and garages could see their performance, behavior changed quickly. The dashboards shifted conversations from opinion to evidence. ## Conclusion & Next Steps The transformation delivered more than a new platform; it gave the organization a scalable, data-backed operating model. By unifying catalog ingestion, compatibility checks, quoting, and installation workflows, the network can now grow without adding equivalent operational headcount. Next steps include expanding predictive inventory insights, automating vendor pricing recommendations, and integrating a customer self-service portal for repeat orders. With the foundation in place, the company is positioned to grow its regional footprint while maintaining consistent quality and responsiveness.