Receiving Defects Aren't the Inspector's Fault
A distribution center receives 2,500 shipments per month across two receiving lanes. Last month it recorded 14 defects: six wrong quantities, four damaged cartons, three missing documents, and one wrong SKU. Each shipment has five critical checkpoints. Receiving cycle time averages 42 minutes against a 30-minute standard, overtime is 12%, and the 5S audit score is 68%. Standard work exists but is not followed consistently. Two shifts operate with six receivers; the morning shift handles 60% of volume. Customer complaints about shorts have risen, and the DC manager wants a quick fix. The quality team has one Black Belt available for eight weeks and a limited budget for scanning hardware. What should the improvement team do?
1. Write a DMAIC problem statement and define the primary metric. Include the baseline DPMO and yield.
Read the worked answer
Problem statement: between the last two monthly audits, the receiving process handled 2,500 shipments with 14 recorded defects, average cycle time of 42 minutes versus a 30-minute standard, 12% overtime, and a 5S score of 68%; standard work is not consistently followed. The problem is measurable and not a solution statement.
Primary metric: defects per million opportunities at the five critical checkpoints. DPMO = 14 / (2,500 x 5) x 1,000,000 = 1,120. Checkpoint yield = (12,500 - 14) / 12,500 = 99.888%. Shipment-level defect rate is 14/2,500 = 0.56%, which is easier for operations to understand. Secondary metrics: cycle time, overtime, 5S score, and standard-work audit compliance. Baseline all five before the pilot so improvement is not claimed from opinion.
2. Use the defect data to prioritize root causes and design a 30-day pilot.
Read the worked answer
Pareto: wrong quantity 6 (43%), damage 4 (29%), missing documents 3 (21%), wrong SKU 1 (7%). Focus on wrong quantity and damage, which together are 71% of defects. Likely causes: no scan verification at put-away, mixed cartons on the lane, congestion at the 42-minute cycle time, and standard work not followed when the morning shift runs 60% of volume.
Pilot for 30 days on one lane: add scan verification for quantity at receipt, separate damaged cartons immediately, use a visual lane layout and 5S reset, and audit standard work twice per shift. Do not buy scanning hardware for both lanes before the pilot. Measure DPMO, cycle time, overtime, and audit compliance weekly against the untouched lane. This tests the cause-and-effect relationship while limiting cost.
3. What control plan and validation would you require before scaling?
Read the worked answer
Validate with a before-after comparison on the pilot lane versus the control lane, using DPMO, cycle time, and overtime. Require at least four weeks of stable data and no evidence that defects were pushed downstream. The control plan should include a daily DPMO chart, a standard-work audit, a 5S checklist, a scan-exception log, and a named response owner for each out-of-control signal. Train the second shift only after the pilot shows a sustained drop.
Scale in stages: first lane, then both lanes, then add hardware if the scan data prove the need. Build a poka-yoke where possible, such as scanner confirmation before put-away, rather than relying on reminder training. Close the project only when the gain holds for 30 days after the Black Belt leaves.