Every Velocity Cycles bicycle is built to a single customer's specification. Custom frames, components selected to the rider's preference, finishes chosen from a catalog that runs into the hundreds of combinations. That level of customization means no two builds share an identical parts list, and no inventory prediction model works cleanly when the inputs change with every order. Velocity was managing that complexity in spreadsheets, and the spreadsheets could not keep pace.

Assembly line bottlenecks came from a predictable cause: a single missing component stops production on that specific bike entirely. In batch manufacturing, a shortage delays a production run and other runs continue. In custom single-unit manufacturing, a missing part delays a specific customer's order for as long as it takes to source what is missing. Velocity was discovering those shortages at the assembly station, not upstream where there was still time to act.

The financial picture was equally fragmented. Labor costs were tracked separately from materials. Shipping charges from international suppliers were not consistently attributed to specific builds. Overhead was distributed as an estimate rather than a tracked cost. The actual margin on any given bike was not knowable until well after it shipped, if at all. Pricing decisions were based on historical averages rather than the confirmed cost of each specific configuration.

The deployment covered per-component inventory tracking from supplier delivery through each assembly stage, real-time assembly progress monitoring per build, cost consolidation across labor, materials, shipping, and overhead per bike, customer order integration with specification and delivery tracking, demand forecasting and scenario planning for parts procurement, and assembly tool maintenance scheduling. Configuration was completed in weeks, not months.

Assembly delays from missing components dropped after FireFlight went live. The change was structural: shortages that had previously been discovered at the assembly station were now visible upstream, when there was still time to expedite, substitute, or adjust the delivery commitment. The delivery failure mode did not disappear overnight, but its cause shifted from discovery to decision. The team had options instead of a stopped production line.

Per-bike profitability became real data for the first time. Configurations that had appeared profitable based on average cost estimates were confirmed. Others revealed that international component sourcing, special finish processing, or labor-intensive assembly stages were consuming more margin than the pricing model had assumed. Velocity's pricing decisions moved from historical averages to confirmed per-unit costs.

When every unit is unique, the cost attribution problem is more acute than in production-run manufacturing. A production run of identical items can absorb cost averaging across the batch. A single custom bicycle cannot. The actual cost of that bike is what was spent on that specific frame, those specific components, that specific finish, and the labor hours that built it. Without per-unit cost tracking, pricing for custom products is based on estimation rather than measurement.

The second thing this deployment confirmed is that demand forecasting works even when every configuration is unique. The components that Velocity used most frequently across their custom builds, specific bearing types, drivetrain families, and tubing sizes, appeared in enough configurations that historical usage patterns were meaningful predictors of future demand. FireFlight identified those patterns and put procurement ahead of the build schedule for the first time.

Deployments for custom build-to-order manufacturing covering per-unit cost tracking, assembly stage monitoring, and supplier lead time management are completed in weeks, not months. The configuration built for Velocity Cycles applies directly to any manufacturer building unique products to customer specifications from a multi-supplier parts base.