Takt, the Constraint, and a Manager's Wish List
A packaging line supplies 480 cases per shift to a customer on daily delivery. Net available time is 450 minutes per shift after breaks and planned maintenance. Five stations run in series with cycle times of 48, 52, 61, 40 and 45 seconds. Station 3, the case erector, is the slowest and also stops for a 20-minute changeover three times per shift because three case sizes run every shift. About 300 cases of WIP sit between stations, and the line is scheduled as a batch push from a weekly plan. Station 4 has a 4% rework loop adding 20 seconds per reworked case, on top of its 40-second cycle. Final-test first-pass yield is 91%. Line overtime last month was 96 hours. The plant manager's improvement list, in order, is: (1) buy a faster case packer for station 5, (2) run a 5S event in the warehouse, (3) add a third shift.
1. Calculate takt time, identify the constraint, and quantify the capacity gap. Then evaluate item (1) on the manager's list.
Read the worked answer
Takt time = 450 min × 60 s / 480 cases = 27,000 / 480 = 56.25 s per case. Every station must complete a case in 56.25 s or less to meet demand in available time.
Constraint: station 3 at 61 s exceeds takt. Gross capacity = 27,000 / 61 = 442.6 cases/shift. Subtract the three 20-minute changeovers (60 min): 390 min × 60 / 61 = 383.6 cases/shift. Shortfall = 480 − 383.6 = 96.4 cases, about 98 minutes of running at 61 s every shift - which is exactly what the 96 hours of overtime are buying.
Item (1), a faster station 5: station 5 runs at 45 s, giving 27,000/45 = 600 cases/shift of capacity against a 480-case demand. It already has 25% slack. Speeding it up raises the capacity of a station that is never the limit, so line output does not change by a single case; you would simply build WIP faster in front of the next constraint. It is the classic 'improve where it is easy' error.
Note also that station 4's effective load is 40 + 0.04 × 20 = 40.8 s, still well under takt. Station 3 is the only station above takt.
2. Rank the available actions and recommend a sequence for the next 90 days, including your view on items (2) and (3). Justify each choice against the constraint.
Read the worked answer
Principle: throughput only rises if the constraint gains capacity or loses lost time. Everything else is preparation or waste removal.
Priority 1 - attack changeover at station 3 (SMED). This is the cheapest capacity on the table. Cutting 20 min to 10 min recovers 30 min per shift = 30 × 60/61 ≈ 29.5 cases/shift, and a 5-minute target recovers about 44 cases. Observe the changeover, separate internal from external work, pre-stage tooling and cases, use quick clamps and a checklist. Cost is hours, not capital.
Priority 2 - reduce station 3 cycle time below takt. Target ≤ 56 s through method work, feeder improvements and eliminating minor stops. This is what converts 383 cases into 442+, and with changeover gains, into 480 without overtime.
Priority 3 - rework at station 4. At 4% it adds 0.8 s average load plus disruption and scrap; fix the cause of rework because quality losses at any station still consume constraint-paced time downstream.
Item (2), 5S in the warehouse: keep it only if downtime data show line-side material starvation as a real loss at station 3. Otherwise it is comfort work that will not move output - park it.
Item (3), third shift: it would meet demand but institutionalizes the constraint, adds headcount and hides the changeover waste. Use it as a temporary bridge or for demand growth, not as the improvement.
WIP: cap it between stations; 300 cases of batch-push WIP hides station 3 stops and lengthens lead time.
3. After the improvements, what must go into the control plan so the gains do not erode over the next six months?
Read the worked answer
The gains are behavioural and will decay without structure.
- Standard work at the constraint. A posted station 3 work sequence with the target cycle of ≤ 56 s, plus a timed changeover checklist with a target of ≤ 10 min. The standard is the new baseline, not a suggestion.
- Visual control at takt. An hourly output board by pitch (e.g., 64 cases per hour) that shows planned vs actual and the reason for any miss. Takt-paced boards expose loss immediately; monthly reports do not.
- Metric set with owners. Station 3 cycle time, changeover duration by event, station 4 rework %, first-pass yield, WIP between stations, and overtime hours. Each has a named owner and a review frequency.
- Reaction plan. If station 3 exceeds 56 s for two consecutive hours, or a changeover exceeds 12 minutes, the team leader escalates and records the cause the same shift. Define who responds and what they check.
- WIP caps and pull. Marked maximum levels between stations; stop the upstream station when the cap is reached.
- Audit cadence. Weekly layered audit of the standard work and changeover checklist for the first 8 weeks, then monthly. Record deviations and feed them back into the standard.
- Re-baseline capability and capacity after the changes, so the next improvement cycle starts from measured values rather than estimates.