Productivity and utilization: how much you get from what you have
Utilization, OEE as lost minutes, labour productivity, and why a team that is busy all the time makes customers wait. Examples from a bakery in Egypt and a pharmacy in Singapore.
Key takeaways
- Three numbers tell you how well a business uses what it has: utilization (how much of the time a machine or person is busy), OEE (how much good output a machine makes against what it could make), and labour productivity (output per hour worked).
- Common mistakes: Pushing customer-facing staff towards 100 percent utilization and then wondering why queues explode.
- Utilization = time busy / time available. A delivery van out on the road 6 of its 8 hours is 75 percent utilized.
- Overall equipment effectiveness (OEE) asks: of all the minutes a machine was planned to run, how many minutes made good products at full speed?
- Labour productivity = output / hours worked. The US Bureau of Labor Statistics measures a whole economy this way, as output per hour worked.
Key idea
Three numbers tell you how well a business uses what it has: utilization (how much of the time a machine or person is busy), OEE (how much good output a machine makes against what it could make), and labour productivity (output per hour worked). Higher is usually better, but not always: people who are busy all the time make customers wait.
The three numbers in plain words
- Utilization = time busy / time available. A delivery van out on the road 6 of its 8 hours is 75 percent utilized.
- Overall equipment effectiveness (OEE) asks: of all the minutes a machine was planned to run, how many minutes made good products at full speed? Minutes go missing in three ways: stops (breakdowns, changeovers), slow running, and defects (products that must be thrown away or redone).
- Labour productivity = output / hours worked. The US Bureau of Labor Statistics measures a whole economy this way, as output per hour worked. In a business it might be orders picked per hour, patients seen per doctor per day, or loans processed per staff member.
The industrial manufacturing module works out OEE for a machining line by multiplying three rates. Here is the same idea counted in minutes, which is easier to explain to a client, because each lost minute has a cause someone can fix.
Worked case
Where a bread line in Cairo loses its minutes
The prompt
A bread line at a bakery in Cairo, Egypt is planned to run 600 minutes a day. At full speed it bakes 2 trays a minute (0.5 minutes a tray). Yesterday it stopped for 60 minutes (a breakdown and a change of recipe), baked 900 trays, and threw away 60 burnt or broken trays. What is its OEE, and where do the minutes go? (Fictional bakery, illustrative figures.)
The structure
- OEE = good minutes / planned minutes
- Stops: planned minutes minus running minutes
- Key: Slow running: running minutes minus the minutes the trays should have taken
- Defects: minutes spent on trays that were thrown away
The exhibit
Waterfall chart: From planned minutes to good minutes on the bread line. Values in minutes. Planned minutes, total: 600; Stops, change: -60; Slow running, change: -90; Defects, change: -30; Good minutes, total: 420.
Working it through
1. Running minutes
Planned 600 minus 60 minutes of stops.
Running minutes:600 - 60 = 5402. Minutes the trays should take
900 trays at the ideal 0.5 minutes each.
Ideal minutes for 900 trays:900 × 0.5 = 4503. Slow running
The line ran 540 minutes but did only 450 minutes of full-speed work.
Minutes lost to slow running:540 - 900 × 0.5 = 904. Defects
60 thrown-away trays used 0.5 minutes each.
Minutes lost to defects:60 × 0.5 = 305. Good minutes
840 good trays at 0.5 minutes each.
Good minutes:(900 - 60) × 0.5 = 4206. OEE
Good minutes divided by planned minutes.
OEE (fraction):420 ÷ 600 = 0.77. The same answer from the three rates
Availability 540/600, performance 450/540, quality 840/900, multiplied together.
OEE from the three rates (fraction):(540 ÷ 600) × (450 ÷ 540) × (840 ÷ 900) = 0.7
What the exhibit shows
Slow running, not breakdowns, is the biggest loss: 90 of the 180 lost minutes.
The recommendation
OEE is 70 percent: of 600 planned minutes, only 420 made good bread. Start with slow running, which loses 90 minutes, more than stops (60) or defects (30). Find out why the line runs below 2 trays a minute: dough not ready, an oven set cooler to avoid burning, or staff short at the start of the shift. Then cut the recipe change time. Every minute won back here is extra bread with no new oven.
Why busy people make customers wait
For machines, higher utilization usually means lower cost per unit. For people who serve customers, it has a hidden price: queues. Customers do not arrive evenly, and some take longer than others. When staff are busy almost all the time, there is no spare time to absorb a rush, so waits grow very fast. For one server with random arrivals, queueing theory gives a simple rule: average wait in the queue = service time x utilization / (1 minus utilization). Teams with several servers wait less at the same utilization, but the shape is the same: waits climb steeply as utilization nears 100 percent.
Worked case
A clinic pharmacy in Singapore with one pharmacist
The prompt
A clinic pharmacy in Singapore has one pharmacist. Each prescription takes 4 minutes on average. 12 prescriptions arrive an hour. The clinic expects patients to rise to 13.5 an hour next year. How long will patients wait, and what should the clinic do? (Fictional clinic, illustrative figures.)
The structure
- Wait in queue = service time x utilization / (1 minus utilization)
- Capacity = 60 minutes / minutes per prescription
- Key: Utilization = arrivals / capacity
- Options: cut the service time, or add staff at peaks
Working it through
1. Capacity
60 minutes divided by 4 minutes each.
Prescriptions an hour:60 ÷ 4 = 152. Utilization today
12 arrivals against a capacity of 15.
Utilization today (fraction):12 ÷ 15 = 0.83. Wait today
4 minutes x 0.8 / 0.2.
Average wait today (minutes):4 × 0.8 ÷ (1 - 0.8) = 164. Utilization next year
13.5 arrivals against 15.
Utilization next year (fraction):13.5 ÷ 15 = 0.95. Wait next year
4 minutes x 0.9 / 0.1.
Average wait next year (minutes):4 × 0.9 ÷ (1 - 0.9) = 366. With a technician who does the labelling
Service time falls to 3 minutes, so capacity is 20 an hour and utilization 13.5 / 20.
Average wait with a technician (minutes):3 × (13.5 ÷ 20) ÷ (1 - 13.5 ÷ 20) = 6.23
The recommendation
A rise of 12.5 percent in patients more than doubles the average wait, from 16 to 36 minutes, because the pharmacist goes from 80 to 90 percent busy. Adding a technician to label and pack medicines cuts the service time to 3 minutes and the wait to about 6 minutes. Do not judge this team by how busy it is: a pharmacist who is 90 percent utilized looks efficient on paper while patients wait over half an hour.
Risks: The rule gives averages; the lunchtime peak will be worse; A technician costs money every hour, so check the cost against the value of shorter waits.
Two warehouses of a retailer pick customer orders. Warehouse A picks 60,000 order lines a week using 1,200 labour hours. Warehouse B picks 45,000 lines using 1,000 hours. How many more lines per labour hour does A pick than B?
Pushing customer-facing staff towards 100 percent utilization and then wondering why queues explode. Quoting one OEE figure without saying which loss (stops, slow running or defects) is the biggest. Comparing productivity between sites that do different work. Calling a machine 95 percent utilized when half its output is defects: busy is not the same as useful.
A machine has OEE of 60 percent. What is the most useful next step?
A one-person help desk goes from 80 to 95 percent utilization. What happens to the average wait?
A team handles the same number of orders with 10 percent fewer hours. What happened to its labour productivity?
Airlines, hotels, hospitals and factories earn their profit by filling costly assets. The "Fill the assets" pattern shows why the last few customers are so profitable.
Open the "Fill the assets" patternSources for this lesson (2)
- Recognized public explanations of case-interview concepts and terms
- US Bureau of Labor Statistics, Productivity (labor productivity measured as output per hour worked) (checked 2026-10-01)
My notes on this lesson
0 of 5,000 characters. Saves automatically.
Try the 4 remaining checks and drills above to complete this lesson (0 of 4 done).
Spotted something wrong or out of date? Report a mistake. We check every report and correct the page.