Batch Size and Changeover at Northline Valves
Northline Valves makes industrial ball valves in a single plant. Customer lead time is six weeks, on-time delivery is 92%, and there are 4,200 finished and in-process valves on the floor. The main machine cell runs batches of 500 parts and takes four hours to change over. Sales forecasts a 10% volume increase next year but the plant has no capital budget for another machining center. The operations manager proposes a second shift and larger batches of 800 to 'absorb' changeover and improve machine utilization. The lean coordinator argues for smaller batches and faster changeover instead. The plant manager wants a decision this week and asks for a plan that protects service without adding people or equipment. Recent quality data show two customer escapes traced to parts waiting three weeks between machining and assembly.
1. Using lean principles, evaluate the two proposals. What should the plant do first, and what data would you use to decide?
Read the worked answer
First separate the proposals by the problem they solve. Larger batches improve reported machine utilization by spreading four hours of changeover over 800 parts instead of 500, but they increase WIP and lead time. The plant already has 4,200 valves in process and a six-week lead time; larger batches make the three-week wait between machining and assembly worse. A second shift may add capacity, but it adds supervision and handoff risk and does not address the changeover constraint.
The first action is to measure the bottleneck and calculate EPEI: available time at the constraint divided by average daily demand for each part family. If the constraint runs 16 hours and daily demand for a family is 80, EPEI is 12 minutes; a four-hour changeover cannot support that, so SMED is the first lever, not batch size. Also pull WIP down between machining and assembly, set a kanban cap, and track lead time, on-time delivery, WIP, and changeover time. Do not promise smaller batches before changeover and quality are stable.
2. The operations manager says smaller batches will reduce machine utilization and increase cost per part. How should the lean coordinator respond, and what risks must be managed?
Read the worked answer
The coordinator should agree that utilization matters at the constraint, but challenge the cost-per-part calculation. Four hours per 500 parts is 0.48 minutes per part; per 800 it is 0.30 minutes, but that accounting ignores inventory carrying cost, expediting, quality risk, and lost responsiveness. If SMED cuts changeover to 90 minutes, a 100-part batch has 0.9 minutes per part while cutting lead time and WIP. Use total cost and flow, not local machine cost.
Risks include starving downstream during changeover, inconsistent quality after frequent setups, and supplier variability. Mitigate with standard changeover checklists, first-piece approval, a buffer or kanban only at the constraint, and a pilot on one part family. Measure changeover time, first-pass yield, WIP, and on-time delivery before scaling.