The Shift You Added Was Already in the Building
By Sameer P, Founder, Sarga II
The Shift You Added Was Already in the Building
A recent LinkedIn post described a manufacturer that added a second shift to meet real demand, only to discover that the first shift had been running at 61% of its rated hourly output. The author pointed to long changeovers, a slow morning startup, a quality check that stopped the line, and repeated walks to retrieve material. Six months later, the second shift was removed and the first shift was reportedly running at 84%, with capacity still available.
“They added a second shift to handle demand. Demand was real. The shift was not the answer.”
“They added a second shift to handle demand. Demand was real. The shift was not the answer.”
That account is not an audited study, and its figures should not be treated as a typical plant result. But the pattern is familiar: a missed output target appears to call for more hours, people, or equipment before anyone has measured how much productive time the current system actually converts into good output.
The distinction matters. Demand may be real while the proposed capacity solution is still wrong. When a line loses time to avoidable stops and uneven flow, adding a shift can increase payroll without addressing the constraint that keeps orders late.
When a Capacity Problem Is Not a Machine Problem
Rated capacity is a technical ceiling under defined conditions. Effective capacity is what the operating system can repeatedly deliver with its actual product mix, staffing, material presentation, maintenance windows, quality checks, and changeovers.
Those numbers are often treated as if they were interchangeable. A plant sees backlog rising and assumes it needs more hours. A sales commitment is translated into a target. The target is compared with machine specifications, not with a measured account of where time disappears during the shift.
This skips the diagnostic question: what is preventing good product from leaving the system at the required rate? A machine can be available while its operator waits for material. A fast upstream process can keep running while work piles up before a slower downstream step. An inspection can protect quality and still interrupt flow if the handoff, staffing, or timing is poorly designed.
Adding a shift may be the right answer. It is a costly answer to choose before the current losses are understood.
Root Cause: The Plant Measures Output, Not Lost Time
Many plants know yesterday’s production total and the machine’s nominal cycle time. Fewer can explain, by shift and product family, where the gap between expected and actual output came from.
A single daily total hides different losses. A startup delay looks like low performance. A changeover may be recorded as planned downtime. A wait for material can be coded as operator delay. Rework can appear as extra production activity even though it consumes capacity without creating another saleable unit.
When these categories are inconsistent, leaders compare lines and shifts using numbers that do not describe the same thing. They may respond to a low-output area with overtime, even though its largest losses are a recurring setup problem or a quality hold.
Core Insight: If the loss is not visible in a shared time account, the next capacity decision is mostly guesswork.
Root Cause: Changeovers and Startups Are Treated as Background
A changeover is easy to accept as an unavoidable part of production. A startup loss can become “how mornings work.” Once these losses are normalized, they stop appearing as opportunities to recover capacity.
In the LinkedIn account, the author reported a 90-minute changeover against a 20-minute target, plus 45 minutes lost at startup. Those figures may be specific to that operation, but they show why a shift average can mislead. If the line has only one such changeover and one such startup in an eight-hour shift, the reported excess is 115 minutes, almost 24% of scheduled shift time. That arithmetic uses the author’s account; it is not an independently audited measurement.
The first task is not to demand that operators move faster. It is to observe the work, separate internal setup from work that can happen while equipment is running, check whether tools and materials are ready, and determine what causes the startup delay.
Core Insight: A loss repeated every shift is not a one-off disruption; it is part of the process design.
Root Cause: Local Efficiency Hides the System Constraint
A line is a connected flow, not a collection of independent machines. Improving a fast upstream station does not increase finished output if a downstream process cannot absorb the extra work. It can instead create more work in process, more handling, and longer queues.
A separate trade-press case described a solar module line where a downstream transfer element constrained output. Synchronizing material movement reportedly raised capacity from about 200 to 215 modules per hour. The author also reported lower recorded downtime and improved runtime after the adjustment, while describing other quality improvements made without added equipment or headcount. These are the author’s case figures, not a general forecast for other factories.
The lesson is not that every line has the same bottleneck. It is that local utilization can look healthy while system throughput remains constrained. If measures reward keeping each machine busy, upstream teams can flood the real constraint and make delivery performance worse.
Core Insight: Improve the step that limits good output, not the step that is easiest to make look busy.
The Real Cost Is More Than Another Paycheque
An added shift brings direct labour expense, but its full cost can include supervision, training, maintenance coverage, utilities, quality support, and the coordination burden of running across more hours. If the new shift is not consistently supplied with work, qualified people, and equipment in good condition, paid time may not become saleable output.
The cost of the existing losses also compounds. Time spent in changeover, startup, waiting, walking, inspection queues, and rework is capacity that cannot be used for another order. If demand is already constrained, the plant may pay overtime or expedite work while still missing promised dates.
A practical estimate should use local data: calculate the recurring lost minutes by product and shift; convert only the recoverable portion into good units; then multiply those units by contribution margin. Add avoidable overtime, premium freight, scrap, and rework where evidence supports them. Compare that total with the fully loaded cost of the proposed shift.
Do not claim every recovered minute becomes output. Demand, materials, labour skills, and downstream capacity may limit the conversion. The point is to make those assumptions explicit before committing cost.
The Fix: Measure the Gap Before Buying Capacity
Sarga II’s diagnostic approach starts with a short, direct observation of the product families and shifts driving the backlog. Establish the required rate, then map actual time from shift start through good output. Use consistent categories for running, setup, planned stops, unplanned downtime, waiting for material, quality checks, and rework.
Next, identify the constraint in the full route. Follow work from release to completion, note where queues build, and check whether the constraint has the material, staffing, tooling, and maintenance support it needs. Compare product mix and changeover sequence, rather than treating all hours as equal.
Choose one recurring loss with a clear owner and a testable countermeasure. Run the change in one area, on a representative product mix, and define the measures in advance: good units per staffed hour, schedule attainment, setup duration, first-pass yield, or queue size. Keep safety and quality conditions unchanged.
Then review the results against the baseline and check whether the improvement holds across shifts. If the loss moves elsewhere, the intervention has not yet improved system throughput.
A Case Pattern: The Second Shift That Did Not Stick
In the reported LinkedIn case, demand had increased, so leadership added a second shift. The first shift’s reported output was below its rated rate. The post described a long setup, lost startup time, an interrupting quality check, and repeated material walks. After the plant addressed performance on the first shift, the author said the second shift was withdrawn and spare capacity remained.
The details are not sufficient to reconstruct the company’s full economics, product mix, or measurement method. We should not conclude from one account that second shifts are generally a mistake. The useful pattern is the sequence: a capacity decision was made before the existing shift’s loss structure was understood; after the losses became visible, the need for added hours changed.
A second shift can still be justified when demand is durable, the constraint is already operating effectively, and the incremental output supports the cost. The diagnostic protects that decision by showing what additional hours can realistically produce.
What Good Looks Like
A plant that understands its capacity can explain the difference between rated output, demonstrated output, and the output it can reliably promise. Supervisors spend less time compensating for preventable shortages and more time managing genuine exceptions. Planners can see the effect of product mix and changeovers. Maintenance, quality, and production agree on what counts as lost time and who owns the next action.
The constraint has a defined buffer and a stable supply of qualified labour, materials, and tools. Upstream work is released at a rate the downstream system can absorb. Improvement tests are tied to finished output and delivery, not just machine utilization.
Most importantly, management can distinguish recoverable capacity from theoretical capacity. If the analysis shows that the line is already performing near its practical limit, the case for another shift or equipment investment becomes stronger, not weaker. It is supported by observed demand, constraint data, and a clear estimate of incremental good output.
If This Pattern Is Familiar
Before approving another shift, ask what the current shifts lose to setup, startup, waiting, quality interruption, and rework, and whether those losses have been measured consistently. Then ask which operation actually limits finished output.
The answer may still be to add capacity. It may also reveal output that is already being paid for but never reaches the customer.
If this pattern is familiar, we should talk. Sarga II: https://www.sarga-ii.com
Comments