A Pilot Under Peak-Season Pressure in a Grocery Distribution Center
A regional grocery distributor's 24-aisle distribution center handles 11,000 SKUs and 38,000 lines per day. Pickers average 85 lines per hour against a target of 110; overtime is 12% and mispicks are 0.9 per 1,000 lines. Travel is 6.2 miles per picker per shift. The operations director proposes a $2.1 million goods-to-person automation system with a 3.5-year payback. The CI manager has $150,000 and a 10-week window before peak season. Data show 8% of SKUs generate 55% of lines, but they are scattered across all aisles; the top 200 SKUs are placed without regard to velocity. Pickers use paper lists and a tall cart. A one-zone pilot is possible in aisles 1-6, which contain 22% of lines. Supervisors fear productivity will drop during the pilot and want to postpone until after peak.
1. Assess the automation proposal and recommend an alternative improvement portfolio within the $150,000 and 10-week constraint.
Read the worked answer
The automation proposal may be strategically attractive, but it cannot be validated within 10 weeks and its payback is based on the current low productivity of 85 lines per hour. If Lean improvements raise productivity toward 110, the labor savings and payback change materially. A $2.1 million commitment before testing the process is a high-risk response to a travel and slotting problem.
Within the constraint, use a portfolio: ABC velocity slotting so high-frequency SKUs sit in the golden zone; batch picking of four to six orders to reduce trips; a redesigned low-cost cart with order compartments; pick-path sequencing; and standard work for cart setup and replenishment. Because 8% of SKUs generate 55% of lines, slotting has high leverage. Aisles 1-6 contain 22% of lines, making them a suitable pilot zone. Reserve some budget for WMS configuration and labels. Do not spend the full $150,000 before pilot data show travel reduction and no mispick increase.
2. Design the pilot in aisles 1-6. Specify metrics, comparison, duration, success criteria, and how to protect peak season.
Read the worked answer
Use a four-phase design: two weeks of baseline in aisles 1-6 and a comparison zone, four weeks of pilot, one week of analysis, and three weeks of buffer before peak. The comparison zone should be similar in lines per aisle and SKU mix, for example aisles 7-12. Keep the demand mix and shift schedule as stable as possible.
Metrics: lines per hour, travel distance or time per pick, mispicks per 1,000 lines, overtime hours, safety incidents, and replenishment delays. Predefine success: at least a 15% pick-rate gain, mispicks no worse than 0.9 per 1,000, and no safety deterioration. Include a stop rule: pause if mispicks exceed 1.2 per 1,000 or pick rate falls below 75 for two consecutive days. Use daily huddles and a simple visual board. If peak arrives early, freeze the pilot and revert to the standard layout; do not run an uncontrolled experiment during the highest-volume period.
3. How would you address supervisor resistance and the concern that the pilot will hurt productivity before peak?
Read the worked answer
Treat the supervisors as co-designers, not obstacles. Ask them which aisles and shifts are least risky, involve them in slotting decisions, and agree on the stop rule in advance. Acknowledge the real concern: a learning curve will temporarily slow picking. Budget temporary labor for the first three days and schedule the pilot away from the highest-volume weeks.
Explain the WIIFM in their terms: less travel, fewer hot-order exceptions, and a credible basis for deciding whether automation is needed. Share daily data rather than defending the pilot. Address fear of layoffs if that is present; if productivity gains will reduce overtime rather than headcount, say so explicitly. Use the sponsor to reinforce that the pilot is time-boxed and reversible. Finally, keep the comparison zone visible so supervisors can see whether the pilot is truly worse or whether the whole DC had a bad week.