Reorder Point and Safety Stock (Service Level)
The trigger level, and why 99% service costs double the buffer.
Example
You enter
- Average daily demand (units/day) 100
- Supplier lead time (days) 7
- Daily demand std deviation (units/day) 20
- Target service level (%) 95
You get
- Service-level z score 1.645
- Safety stock 87
- Reorder point 787
Details, formula, and sources
reorder_point = demand during the lead time (avg_daily x lead_time) + safety stock, where safety_stock = z x demand_sd x sqrt(lead_time) and z is the service-level z-score (95% = 1.645, 99% = 2.326). 100 units/day, 7-day lead, sd 20, 95% service -> 87 units safety stock, reorder at 787; raise to 99% (z 2.326) and it is 123 units of buffer, reorder at 823 -- 36 more units for the last 4 points of service. Safety stock scales with sqrt(lead time) and the z-score, so chasing the last points is expensive. Normal demand, fixed lead time. A planning aid; the demand pattern and supplier reliability govern.
z = inverse_normal(service_level / 100); safety_stock = z x demand_sd x sqrt(lead_time_days); reorder_point = avg_daily_demand x lead_time_days + safety_stock.
Reorder point and safety stock (standard service-level inventory control), first-principles, by name; the actual demand pattern and supplier reliability govern.
The service-level reorder-point and safety-stock relations are standard published inventory-management results; the demand, lead time, and variability come from the business's records.
Estimate. AHJ and licensed professional govern.
Field names used by the API: avg_daily_demand, lead_time_days, demand_sd, service_level_pct, z, safety_stock, reorder_point
- Two parts reorder point = demand during lead time + a safety-stock bufferservice-level inventory control
- Square-root scaling safety stock scales with sqrt(lead time) and the z-score; the last points cost a disproportionate bufferservice-level model
- Fixed lead time assumes normal demand and a fixed lead time; variable lead time adds a second variance termscope of this tile