Naar de inhoud
OneKitly

Seventy Per Cent Capacity Is Not Thirty Per Cent Waste

Gepubliceerd op 23-9-2026 · 3 min leestijd · Zakelijke tools

Lena Hoffmann

Lena Hoffmann — Redacteur wetenschap en onderwijs bij OneKitly

Wiskunde · Natuurkunde

Getoetst aan 2 bronnen

Profiel bekijken →
Kort gezegd

Capacity utilisation is actual output divided by the maximum the plant could sustain: 7,000 against 10,000 is 70 %. The instinct that follows — that the missing 30 % is money left on the floor — is where the reasoning usually goes wrong, because a system pushed towards 100 % does not simply produce more. Queueing theory gives waiting time a term of 1 ÷ (1 − utilisation), which is 3.3 at 70 % and 20 at 95 %: a six-fold rise in queueing for a quarter more throughput. That reserve absorbs the machine that breaks, the order that arrives late and the shift that runs short, and a plant that has spent it has nothing left to absorb anything with. The right question is not how to raise 70 %, but whether the delivery promises the business makes are compatible with the reserve it has chosen to keep.

Advertentie

7,000 units made against a capacity of 10,000 is 70 % utilisation. The reflex is to call the other 30 % a loss; queueing theory says it is the reason the line still delivers on time.

Which capacity is in the denominator

The same plant has at least three capacities: what the machines could do running flat out, what the current shift pattern allows, and what the tightest step in the chain permits. Choosing the first gives a flattering-looking problem — a big gap to close — while the third gives the only number that can be acted on, because a line runs at the speed of its slowest station and every other station's spare capacity is decoration. State which capacity you divided by, or the percentage is not comparable with anyone else's, including your own from last quarter.

Utilisation is not effectiveness

A line can run at 95 % utilisation and produce scrap, or run fast on a product nobody ordered. Utilisation counts hours filled; it says nothing about whether the output was good, needed, or worth making. That is why plants that manage this figure alone tend to accumulate finished goods — the cheapest way to raise utilisation is to keep making something, and inventory is where the something goes. Pair it with a yield or quality measure before letting anyone set a target on it.

Queueing factor
Relative queueing, from the 1 ÷ (1 − u) term
UtilisationOutputQueueing factor
70 %7,0003.3
85 %8,5006.7
95 %9,50020

Uitgewerkt met onze eigen calculator

Calculator capaciteitsbenutting

Gegeven

Werkelijke productie
800
Maximale capaciteit
1.000

Resultaat

Capaciteitsbenutting
80%

Deze cijfers komen uit de calculator hieronder, ze zijn niet met de hand ingetypt — ze worden opnieuw berekend telkens als de tool verandert.

Reken het door met je eigen cijfers →
Advertentie

Veelgestelde vragen

So what utilisation should we aim for?
It depends on how variable the work is, and that is the whole answer. A process with predictable arrivals and identical jobs can sit near 90 % without queues building; a workshop with urgent one-off jobs and unreliable inputs starts choking well before 80 %. Rather than borrow a target, plot your own lead times against your own utilisation over a few months — the point where the curve turns upwards is your ceiling, and it is a fact about your variability, not about the industry.
Does this apply outside manufacturing?
Yes, and the effect is sharper where the capacity is people. Booking a team at 100 % of its hours leaves nothing for the interruption, the sick day or the piece of work that turns out to be twice its estimate, so the plan fails on contact rather than absorbing. The same 1 ÷ (1 − u) shape governs a hospital's beds, a consultancy's billable weeks and a support queue — which is why the well-run ones deliberately plan below full.

Artikelen die je misschien interesseren

Alle gidsen →
UitlegProcess Capability: Cp, Cpk and What Six Sigma Actually ClaimsCp compares the spec width to the process spread; Cpk penalises being off-centre. A process can have an excellent Cp and still make scrap — here is the case, with defect rates computed from the normal distribution rather than read off a table.UitlegTakt Time, Cycle Time and Lead Time Are Three Different ClocksTakt is demand, cycle time is capability, lead time is what the customer experiences. Confusing them is the most common failure in a first improvement project — and Little's Law is the bridge from one to the next.UitlegFive Points of Learning Rate Turn Fifty Repetitions Into Four HundredAt an 85 % learning rate, a task that takes 20 minutes first time reaches 8 minutes after 50 attempts. At 90 % it needs 415 — the same task, five points apart.UitlegA Labour Cost Ratio Means Nothing Until You Say What Is in It$8,000 of wages on $25,000 of revenue is 32 %. Add $1,800 of employer contributions and the same week reads 39.2 % — a seven-point move with no change in staffing.UitlegOutsourcing the Support Desk Raises Revenue per Employee by a Quarter25 million of revenue across 250 people is 100,000 each. Move fifty of them to a supplier and the same revenue reads 125,000 — a 25 % improvement in a productivity measure, with the same work being done.UitlegWhat Is Food Cost Percentage? Portion Drift Beats Supplier PricesA $3.30 plate on a $12.00 menu is 27.5 percent. An extra 30 grams of chicken and 6 percent waste push it to 31.65 percent — while a 5 percent supplier increase moves it only 1.38 points.

Vergelijkbare tools

Bronnen

Een fout in dit artikel gezien?