Bottleneck thinking: what it is and what to do instead

The idea worth keeping

Not as boxes to fill, but as questions that fall out of the maths of the goal:

• How many items are in the system, and how fast do they leave?
• Where do items wait longest?
• Is the slow step inside this process, or does it depend on another department?

Why reaching for it fails in an interview

• It is generic. "Find the bottleneck" is said about every process, and on its own it does not find one.
• It misses the driver that matters. In hospital emergency departments, long waits often come from patients waiting for a ward bed, not from the doctors everyone looks at first.
• It sounds rehearsed. Naming the theory before measuring anything sounds like a slogan.

What to do instead: a worked case

Cutting waits in an emergency department

A hospital's emergency department has an average stay of 5 hours. The target is 4. Where should it act?

The tempting answer: Find the bottleneck: probably not enough doctors, so hire more.

1. 1. Pin the question

   Cut the average time in the department from 5 to 4 hours without adding beds.

2. 2. Write the maths of the goal

   • Average time in the department = patients in the department / patients leaving per hour (Little's law)
   • At 12 patients an hour, a 4-hour stay means about 48 patients in the department, against 60 today

3. 3. Use what you know about the business

   • In hospitals, patients who need admission often wait in the emergency department for a ward bed.
   • Ward beds free up when patients are discharged, and many hospitals discharge late in the afternoon.
   • So the emergency department can be blocked by the wards, a problem more doctors would not fix.

   The structure that falls out of it

   • Patients waiting to be seen
     • Arrivals by hour
     • Doctors on shift by hour
   • Patients waiting for tests
     • Scan and lab turnaround
   • Patients waiting for a ward bed
     • Number waiting
     • Time of ward discharges

   Hypothesis: I expect a large share of the 60 patients are admitted patients waiting for ward beds, because discharges happen late.

4. 4. Find the facts that decide it

   • Today: 60 patients in the department / 12 leaving an hour = 5 hours. Of the 60, 20 are admitted patients waiting for a bed.
   • Moving ward discharges before noon would cut patients waiting for beds from 20 to 8. That leaves 48 in the department: 48 / 12 = 4 hours.

5. 5. Say so what

   Fix the wards, not the emergency department: 20 of the 60 patients are waiting for a bed, and moving discharges before noon would cut that to 8, bringing the average stay to the 4-hour target. Start with a morning discharge plan on the busiest wards. The risk is discharging too fast, so keep the clinical checks and add a discharge lounge.

Why this beats Bottleneck thinking: Little's law shows how many patients must leave the department, and knowing how hospitals work shows that the real constraint sits in the wards.

Build the acumen behind it

The worked case used two things a list cannot give you: the five moves, and knowing how this kind of business makes money. These pages teach both.