Capacity reserve on the shop floor: how to meet deadlines without failures

Why 100% loading breaks deadlines

When a shop is loaded to 100%, the plan looks fine only on paper. In a real shift there are almost always losses: a machine stops, an operator waits for a tool, setup takes longer, the first part goes for recheck. Just one failure can eat up the whole day's buffer.

If a CNC lathe stood idle for 40–50 minutes, you can't place those minutes anywhere. The next job shifts, then another, and by the end of the shift the delay grows on its own. An urgent order doesn’t help either. It gets pushed forward because the client waits, while the shop is scheduled minute-by-minute. As a result, the urgent part is done, but regular orders fall behind.

A common mistake is to count only pure machining time. In practice the shop doesn't operate that way. The operator removes the previous part, sets the fixture, checks the first piece, measures dimensions, sometimes adjusts the program. For small batches and frequent changeovers these operations often take more time than planned.

Full loading removes a shop's simple flexibility. You can’t calmly reschedule work, perform a careful recheck, or insert an unplanned batch without pain. Any small issue immediately impacts promised deadlines.

On a turning shop this is especially visible. Say you planned 8 hours of pure machining on two machines for a day. Then an urgent order for flanges arrives, one setup took 25 minutes longer, and the first-part inspection required a second pass. Formally the deviations are small. In reality, the evening batch moves to the next day.

So 100% loading is not discipline, it's a fragile schedule. A proper plan leaves room for normal shop life: failures, urgent orders, changeovers and inspection. Otherwise even a good shop begins to miss deadlines not because of poor work, but because the schedule is too tight.

What to treat as the shop buffer

The shop buffer isn’t empty hours in the schedule or hidden downtime. It’s part of available capacity that the planner intentionally doesn't allocate to regular orders. It’s reserved for things that almost always happen: minor failures, changeovers, urgent inserts, rework after rejects, waiting for fixtures or tools.

If machines are scheduled at 100%, any delay immediately pushes the queue forward. That’s why the buffer is considered a working reserve, not lost time.

It's better to calculate it not as a single number for the whole shop, but where delays most quickly break delivery: specific machines that already have queues, people who service several machines, and bottlenecks that almost every order passes through.

Keep the reserve where the queue grows fastest. If the turning group overall has free capacity, but one CNC lathe is constantly overloaded, the shop-wide reserve won't help. The buffer is needed on that machine or next to it, where work can be quickly taken over.

This reserve usually covers short stops, tool changes, changeover between batches and urgent orders. Sometimes you also include time for the first-part check if that stage frequently delays the start of a batch.

A simple guideline: the buffer should cover usual week-to-week fluctuations, not a rare two-day breakdown. Major failures are counted separately, otherwise the plan becomes too loose. If a turning shop consistently loses 1–2 hours per shift to changeovers and small failures, it's better to remove those hours from available capacity up front. Then production deadlines will rely on the real situation, not an ideal calendar.

What determines buffer size

You can’t pick buffer size by feel. If you set the same reserve every week, you’re likely to miss. One shop loses time to short stops, another to long changeovers, a third to frequent urgent inserts in an already tight schedule.

Start by looking at losses that repeat almost every week. Usually it’s not big breakdowns but small things: waiting for a tool, program error, fixture change, re-measuring the first part, a pause waiting for a technician. Each might be 10–20 minutes, but over a shift and a week they add up.

Also check how often urgent orders arrive. If the shop takes one urgent part a month, you need a small reserve. If 2–3 urgent items cut into the plan almost every week, the buffer must be larger. Otherwise the planner promises dates based on the regular portfolio, then breaks them with urgent inserts.

Changeovers are also important. One batch might require a 25-minute changeover, and another more than an hour because of different fixtures, chucks, tools and first-part checks. If the week has many small batches with different parts, the buffer should be higher than for long runs.

Don't look only at the month. By month a shop may appear 85% loaded, but within the week one shift may be nearly full while another is free. Deadlines fail not because of total hours, but because the needed machine and the necessary shift are occupied when an order must start.

Another source of delays is neighboring operations. Inspection, washing, marking, packing and shipping often slow production more than machining itself. If the machine finishes a batch on time but quality control picks up parts only at the end of a shift, machine-hour buffers won't save the promised date.

In practice it’s useful to check five things: how many hours the shop actually loses to recurring failures, how often urgent orders arrive, how much time changeovers take for different parts, which shifts are overloaded, and whether queues form after machining. After that, calculate the buffer to match your typical week, not just a precaution.

How to know the buffer is wrong

If the plan looks neat only on paper, the buffer was chosen incorrectly. You notice this not in a pretty chart but in how the shop behaves after normal disturbances: a short changeover, a tool delay, re-measurement of a part or an urgent run.

The first simple sign: the queue grows after every small stop. The machine stood for 20 minutes, and by the end of the shift the shop hasn't caught up. That means there are too few free hours and ordinary time losses immediately push orders to the tail.

The second sign: one urgent order shifts the whole schedule. If the dispatcher inserts an urgent part and every other deadline slides by a day or two, the capacity reserve is too small. Such a plan works only until something unexpected happens.

The third sign: there seems to be a buffer, but it’s in the wrong place. For example, the shop has free hours overall, but an overloaded machine still misses deadlines. Formally a reserve exists, but in practice it’s useless.

There’s also the opposite mistake: free hours remain unused week after week, and deadlines were never at risk. That means the buffer is excessive. It doesn’t help the shop, it just eats available capacity.

A good buffer is usually unnoticeable in a calm week and very noticeable in a bad one. It won’t make the schedule perfect, but it prevents normal disturbances from turning into a chain of delays.

How to set the buffer step by step

You don't choose a capacity reserve by eye. Random 10% can be too big or too small. You need a simple fact-based calculation.

Start with the last month. That’s enough to see recurring losses without drowning in old data that no longer reflects current load.

1. Collect all lost hours by shifts and machines. Take only facts: downtimes, unplanned stops, urgent inserts, extra changeovers.
2. Break those hours down by cause. If everything is mixed into one number the buffer will be inaccurate.
3. Find where delays start most often. It’s not always the most loaded machine. Sometimes one shift, one station, or a group of parts with long changeovers breaks deadlines.
4. Set the reserve by shifts, not for the entire month. If the day shift more often takes urgent orders and stops, it might need 1.5–2 hours of reserve. The night shift may be fine with 30–40 minutes.
5. Test the approach for one week. Compare plan vs actual: how many hours you left free, how many of those hours were actually used for failures, and how many orders shipped on time.

On a turning shop this is convenient because losses are easy to see from downtime and changeover logs. If one CNC lathe disrupts the schedule three times in a week, don’t spread the reserve across the whole shop. Better to leave the buffer where the problem appears first.

After the test week the picture usually becomes clearer. If free hours remain unused, the reserve is too large. If urgent jobs again slipped into promised slots and pushed them out, the reserve is too small. A normal buffer doesn’t have to look good in a table — its job is to hold deadlines in real shifts.

A simple example for a turning shop

Suppose a turning shop has 80 machine-hours per week. The planner has already scheduled five series batches that together take all 80 hours.

On Tuesday an urgent 6-hour order arrives. On Wednesday one CNC lathe stops for 6 hours due to a failure. The shop immediately faces a 12-hour deficit.

Without a reserve

If no capacity reserve was allocated, the planner starts pushing out work that sits later in the week. Usually two regular batches — say 5 and 7 hours — get postponed. The urgent order is inserted, but the problem remains.

Then a chain of shifts follows. The operator changes the sequence, the warehouse waits for parts, the next operation shifts as well. By Friday the shop is living by constant rearrangements, not by the plan.

With a reserve

Now the same shop, but the planner loads the week not to 100% but to 72 hours. Another 8 hours are left as a buffer. This is not empty time but a reserve for failures, urgent orders and small changeover losses.

Again an urgent 6-hour order arrives and a machine stands for 6 hours. The shortage is now 4 hours, not 12. The planner doesn’t break the whole schedule but only moves one low-priority batch whose delivery date has slack.

The urgent order goes into production on time. Two main series batches remain in the week, and the next operation doesn’t receive new shifts. Such a buffer doesn’t make the shop underloaded. It simply buys time when things don’t go as planned.

On CNC lathes this is especially noticeable. Even a short 4–6 hour stop quickly eats a shift if the schedule is nearly full. A good planner watches not only occupied hours, but also which orders can be shifted without harming promised dates.

Common mistakes

The first mistake is simple: planning the shop as if 100% loading is the normal mode. On paper the schedule looks tight. In the shop such a plan quickly cracks: one unplanned setup, one tool delay, one urgent order — and the whole day slides.

The second common error: the planner counts only machine cutting time and forgets that a machine doesn’t cut metal without preparation. Setup, fixture changes, first-part inspection, material delivery, awaiting the program — all of these consume hours. If you don’t include these losses, the reserve will be fictional.

Another weak practice is assigning the same buffer to every machine. It’s convenient, but rarely works. A horizontal center with rare changeovers and a lathe that constantly handles small batches have different rhythms. They need different buffers. Otherwise one machine will be idle while another constantly burns out.

Often teams try to hide longstanding problems inside the buffer that should be fixed separately. For example, a machine regularly fails for the same reason, a technician repeatedly spends an extra hour on a common operation, a program for a typical part requires manual fixes, or blanks arrive damaged or late. That’s not a reserve, it’s hidden losses. If you hide them in the general buffer, the plan will always look acceptable, even though the shop is underperforming.

Another big mistake that hits deadlines hard: promising a date to a client before checking the most loaded stage in the part’s route. An order may have free capacity at the first and last operations, but a turning machine in the middle is booked five days ahead. From the outside the order looks feasible; inside it’s already late.

So it’s better to look not at the average shop load but at the tightest spot. In CNC production one overloaded machine often breaks the promise for the whole order, even if other machines are free.

Quick checks before a new week

Before a new week, spend 10 minutes on a short review. This habit often saves deadlines more effectively than trying to load machines to 100%.

First, check if there are free hours for at least one urgent order. For a turning shop this is usually reserve on the busiest machine, not just the sum of free time across the shop. If the queue sits on one machine, reserve elsewhere won’t help.

Then see if there’s a window for an unplanned stop. A drive failure, long tool change, repeated setup or extra inspection after the first part can easily eat several hours. If there’s no such window, the plan is already too fragile.

Check the busiest machine separately. If it’s nearly at its limit, the whole shop is already running without a buffer, even if other machines are underloaded.

Also verify release against inspection and shipping. Often the shop finishes machining on time, but quality control, packing or transport can’t keep up. Reports look fine, but the client gets a delayed date.

Finally: sales, the foreman and the planner should look at the same plan. If sales promise one date while the shop works to a different sequence, conflicts start Monday morning.

A simple example: if the busiest lathe is booked 36 of 40 hours and you still need 3 hours for a possible stop or urgent insert, there’s almost no free window. Formally the plan fits, but any failure will break deadlines.

A good sign is when you can quickly show where the hour buffer sits, on which machine it is, and who can free it for an urgent job. If no one can identify that at a glance, the buffer exists only on paper.

What to do next

Start with plan vs actual for the last four weeks. Not across the whole plant at once, but for one area where deadlines slip most often. Compare how many hours you promised to load, how many hours machines actually worked, and where losses went: changeovers, downtime, urgent jobs, waiting for tools or material.

Then set a simple loading limit for the bottleneck. If the busiest lathe reliably struggles at 90% planned load, don’t give it 95% just because the chart looks neater. Better to set a clear ceiling, for example 80–85%, and see how the area performs for two to three weeks.

It helps to establish a simple routine: check plan vs actual hours for four weeks, pick one bottleneck and set a loading cap for it, count urgent orders separately from series work, and revisit the buffer after repeated major failures.

Separating series and urgent jobs almost always gives a more honest picture. Series work is steadier and needs a smaller reserve. Urgent orders break the queue, require fast changeovers and quickly eat hours that seemed free on the plan. Mixing these flows into one number makes the buffer overly optimistic.

If the analysis shows the problem isn’t the schedule but the area’s capacity, you can discuss options with EAST CNC. The company supplies CNC lathes and machining centers and helps with selection, commissioning and service.

A good buffer saves deadlines while the shop’s overall capacity is sufficient. When no reserve remains even after proper planning, the problem is not the spreadsheet but the production resource itself.