Inventory

Setting Reorder Points That Survive Real Lead Times

A reorder point is the tripwire that decides whether production runs or waits. This playbook layers lead-time demand, safety stock, and hold reserves, then installs the monthly recalculation that cuts stockouts in half.

The reorder point is the tripwire that decides whether production runs or waits. Set it too low and a $4 fastener stops a $10,000 per hour assembly line; set it too high and you bury cash, because every 1,000 units of unneeded safety stock on a $25 part is $25,000 sleeping on a shelf at a 20 to 25 percent annual carrying cost. Plants with untended reorder points typically carry 20 to 30 percent excess inventory on half their parts while still stocking out on the other half 4 to 8 times a year. One number per part, reviewed on a schedule, decides both failure modes.

Build it in layers: lead-time demand plus safety stock plus minimum order reserve plus quality-hold reserve. Worked example: usage of 150 units a day and a 12 day supplier lead time make lead-time demand 1,800. Safety stock for demand and supply variability adds 400 units, roughly 2.7 days of cover. A minimum order reserve of 200 covers the work cell pulling full totes, and a quality-hold reserve of 150 covers incoming inspection holding 8 percent of lots for an average of 2 days. Reorder point: 2,550 units. The Manufacturing Reorder Point calculator stacks those four terms; your work is keeping each input honest, especially lead time, which drifts quarterly.

Judge the system by two numbers in tension: service level and inventory turns. Target 98 to 99 percent line-item fill on A parts that stop production, 95 percent on B, and 90 percent on C, because a uniform 99 percent target across 8,000 SKUs is a budget-destroying decision made by default. Raw material turns of 10 to 15 are typical for discrete manufacturing; over 20 is excellent; under 6 means reorder points are padded with fear. Also track stockout events per month at the point of use, not just warehouse fill: a plant can report 97 percent warehouse fill while the line starves twice a week on the 40 parts that matter.

Levers that actually move it: first, measure real lead time, receipt date minus order date, on the last 10 purchase orders per supplier; plants find actuals run 20 to 40 percent longer than the item master on a quarter of their parts, and every unmeasured week of lead time is a stockout scheduled in advance. Second, cut supply variability instead of buffering it: a supplier delivering in 10 to 24 days needs triple the safety stock of one delivering in 12 to 14. Third, shrink the quality-hold reserve by fixing incoming rejection rates; taking a supplier from 8 percent of lots held to 2 releases most of that buffer. Fourth, recalculate usage monthly, because a 15 percent demand shift silently invalidates every reorder point set last year.

The failure modes repeat everywhere. Set-and-forget: reorder points calculated once in 2023 still running in 2026, wrong on 60 percent of parts. Double buffering: safety stock in the reorder point, plus planner padding, plus a supervisor's private floor stock, easily 40 percent redundant inventory. Ignoring the quality hold: 300 units on hand, 250 on hold, the system sees no need to order, and the line stops Thursday. Reacting to one stockout by doubling the buffer, which is how a single bad month becomes a permanent $18,000 shelf ornament. And triggering on warehouse balance instead of available balance, so allocated inventory masks the trip.

The management cadence: daily, review parts that tripped below reorder point with no open purchase order, target zero. Weekly: stockout postmortems, where every line-stop part gets a 15 minute review naming which input failed, lead time, usage, safety stock, or hold reserve. Monthly: recalculate reorder points on A parts with trailing 3 month usage and trailing 10 purchase order lead times, flagging every part whose new point differs from the old by more than 20 percent. Quarterly: full B and C recalculation, a dead stock review on parts with no usage in 180 days, and a service-level versus turns report to the plant manager. This rhythm typically cuts stockouts 50 percent in two quarters while lowering inventory 10 to 15 percent.

World-class looks like: 99 percent fill on line-critical parts, raw material turns above 12, stockout-driven line stops under 2 hours a month, and reorder points refreshed monthly by exception with fewer than 5 percent needing manual override. Buyers spend their time on the 30 parts flagged by the recalculation instead of firefighting 300. Lead times in the item master match measured actuals within 10 percent. Quality holds show in available balance within an hour of disposition. And when someone proposes raising a buffer, the first question asked is which input changed, because a reorder point is a calculation, not a negotiation.

Published 2026-07-02.