Case study · Supply chain
Forecasting delivery and procurement from China
A custom pipeline that predicts inbound timing, reorder windows, and out-of-stock risk across Amazon and other marketplaces.
- Client
- Amazon-first seller sourcing from China
- Scope
- Forecasting pipeline + reorder dashboard
- Marketplaces
- Amazon + EU platforms
The problem
The seller sources inventory in China and sells across Amazon and other marketplaces. Every reorder decision depended on a manual spreadsheet ritual: pulling sales velocity, guessing freight times, and hoping the safety stock held.
Two failure modes kept repeating: ordering too late — stockouts and lost ranking, or ordering too early — cash frozen in inventory that sat in a warehouse.
What we built
- A data pipeline that pulls sales, stock, and inbound shipments from marketplace APIs daily.
- A forecasting model for inventory burn rate per SKU, factoring seasonality and promo periods.
- Supplier lead-time and freight-routing profiles: production days, port-to-port, customs, last mile.
- Reorder windows computed per SKU: the latest safe order date, and the cash-optimal one.
- Out-of-stock risk flags with early alerts to the team’s messenger.
How it works day to day
The team opens one dashboard. For each SKU it shows: days of stock left, inbound shipments with predicted arrival, the recommended reorder date, and the risk flag. When a shipment slips or sales spike, the dates recalculate and the responsible person gets a ping.
Outcome
Reorder decisions moved from a day of spreadsheet work to a dashboard the team checks in minutes. Procurement now argues about strategy, not about whose numbers are right.
The pipeline is the client’s property and runs on their infrastructure.
Have a similar problem?
Tell us about your operations.
Describe the process that hurts — we reply within one business day with a clear yes/no and next steps.
- Reply within 1 business day
- Honest go / no-go answer
- No “boilerplate proposal”