Led the 0-to-1 build of Segway's dealer commerce platform, converting a fully manual wholesale operation into a digitized order-to-cash system for 150+ SMB dealers.
When I joined Segway as PM and Sales System Admin in February 2022, the entire dealer wholesale operation ran on phone calls, emails, and Excel spreadsheets. There was no system of record, no order tracking, and no visibility into the pipeline.
I led the 0-to-1 build of the Dealer Portal — a self-service platform that gave 150+ SMB dealers direct access to ordering, inventory, invoicing, and fulfillment tracking. This was my product from day one: I owned the vision, roadmap, and execution.
I was the sole PM on this, reporting to the VP of Sales. The build team was small — one product designer, two contract engineers, and me. On the Segway side, I coordinated daily with Sales Ops, Logistics, and Finance. There was no digital product org when I started.
My responsibilities split into three buckets: writing requirements and specs, managing SAP system configuration and data integrity, and running the dealer rollout. I wasn't just handing off PRDs — I was debugging data sync issues and on calls with dealers walking them through the platform.
The existing order-to-cash process was entirely manual, fragmented across five teams, and had zero digital tracking. Every purchase order touched at least four people before it shipped — and no one had end-to-end visibility into where an order stood at any given moment.
No PO status visibility
No real-time inventory
80/20 operational drag
I mapped the entire order-to-cash workflow before building anything. Every step below was a real handoff point where data was re-entered, context was lost, or errors were introduced.
The Dealer Portal consolidated the entire order lifecycle into a single self-service platform. Dealers could browse inventory, place orders, track fulfillment, and manage invoices — all without picking up the phone or sending an email.
We evaluated Shopify Plus, BigCommerce B2B, and OroCommerce. None of them could handle our SAP integration requirements, dealer-specific pricing tiers, or credit term workflows without heavy customization. Building custom gave us full control over the data layer and let us ship faster than configuring an off-the-shelf tool would have.
Instead of bolting on separate tools for inventory, ordering, and invoicing, we built a unified data layer that synced with SAP in real time. This meant a single source of truth for product availability, pricing, and order status — eliminating the reconciliation work that consumed hours every week.
We deliberately scoped analytics out of the first release. The core problem was that dealers couldn't place orders without calling someone. Solving that came first. Dashboards and reporting came in Phase 3 after we had real usage data to build against.
The initial rollout plan assumed that once we launched the portal, dealers would naturally migrate from the old email-and-phone workflow. That didn't happen. Many dealers had years-long relationships with their sales reps and saw no reason to change their process — even when the new one was objectively faster. We had to redesign the rollout to create real incentives: portal orders shipped first, got priority support, and had guaranteed 24-hour fulfillment. That changed the math for dealers and drove adoption from 30% to 80% in the next 60 days.
A better product doesn’t drive adoption — a better outcome does. We stopped selling the portal and started selling faster shipping.
Phased rollout. Parallel migration. Measured adoption. We didn't flip a switch — we moved systematically.
Started with top 20 highest-volume dealers, then expanded to the full 150+ network over 3 months
Not training videos — actual calls walking each dealer through their first orders on the platform
Both systems ran simultaneously for 30 days. Portal orders shipped first — creating a real incentive to switch
Target: 80% adoption in 60 days, <2% error rate on self-service orders
The hardest part wasn’t building the platform — it was getting 150 dealers to trust a screen over a phone call. Adoption only worked when the transition felt like an upgrade, not a mandate.
I would have invested in an analytics layer from the start, even a lightweight one. We made roadmap decisions in Phase 2 and 3 based on qualitative feedback because we didn't have usage data. That worked, but it was slower than it needed to be. I'd also push harder for a proper staging environment early on — we burned time debugging data sync issues in production that a staging setup would have caught.