Capacity#3

Bottleneck Math: Your System Can Only Move as Fast as Its Slowest Step

Friday, February 20, 2026

Here's a truth that surprises most managers: Adding capacity anywhere except the bottleneck is a waste of money.

Double your packaging speed? Great — but if your assembly line can only produce 50 units/hour, you'll just have faster packaging... of 50 units.

Hire three more customer service reps? Wonderful — but if your technical team can only resolve 20 tickets/hour, you've just created faster intake... of 20 tickets.

This is the core insight of bottleneck math. Master it, and you'll never waste capacity investment again.

The Fundamental Rule

System throughput = Bottleneck throughput.

That's it. Your entire operation can only output what your slowest step can handle. Everything before the bottleneck creates inventory. Everything after it sits idle.

This sounds obvious. Yet most improvement efforts ignore it entirely.

Finding Your Bottleneck (The Quick Method)

Walk your process. Look for these signals:

  1. Pile-ups. Work-in-progress accumulating in front of a station.
  2. Waiting. Downstream resources idle, waiting for input.
  3. Overtime. One step consistently running longer hours than others.
  4. Complaints. The step everyone blames for delays.

The bottleneck is usually where WIP is piling up — not where people are complaining the loudest.

Bottleneck Math in Action

Example: A Manufacturing Cell

| Step | Cycle Time | Capacity/Hour | |------|-----------|---------------| | Cutting | 2 min | 30 units | | Assembly | 4 min | 15 units | | Painting | 1.5 min | 40 units | | Packaging | 1 min | 60 units |

System throughput: 15 units/hour (limited by Assembly)

Now, imagine you're told to "improve the process." Where do you focus?

Wrong answer: Speed up Packaging (already at 60/hr)
Wrong answer: Reduce Cutting time (already at 30/hr)
Right answer: Attack Assembly

If you cut Assembly time from 4 minutes to 3 minutes:

  • New Assembly capacity: 20 units/hour
  • New system throughput: 20 units/hour (now limited by Cutting)

You just increased output by 33% by improving one step.

The Trap: Moving Bottlenecks

Here's what trips people up: when you fix one bottleneck, another appears.

In our example, improving Assembly to 20/hr makes Cutting the new bottleneck (30/hr). Now that's where you focus.

This is why continuous improvement is continuous. You're always chasing the constraint — but each chase moves you forward.

Bottleneck Utilization: The Hidden Metric

Your bottleneck should run at 100% utilization. Always.

  • Every minute the bottleneck is idle = output you'll never recover
  • Lunch breaks, shift changes, maintenance stops — they all cost throughput

This is why smart operations:

  • Stagger breaks around the bottleneck
  • Build buffer inventory before the bottleneck (so it never starves)
  • Schedule preventive maintenance during off-hours

What About Variable Work?

Real processes aren't perfectly predictable. Cycle times vary. Arrivals are random. Equipment fails.

This is where simulation becomes essential.

In a spreadsheet, Cutting averages 2 minutes, so capacity is 30/hr. Simple.

In reality? Some parts take 1.5 minutes. Others take 3. Sometimes the machine jams. The actual throughput might be 25/hr — or 22, or 28.

Simulation captures this variability. It shows you not just the theoretical bottleneck, but the practical one — the step that causes delays most often, under real conditions.

The 5-Step Improvement Process

  1. Identify the bottleneck (look for WIP accumulation)
  2. Exploit it (maximize utilization — no idle time)
  3. Subordinate everything else (non-bottlenecks pace to the bottleneck)
  4. Elevate if needed (add capacity only at the bottleneck)
  5. Repeat (find the new bottleneck)

This is the Theory of Constraints in five steps. Deceptively simple. Devastatingly effective.


Today's Action: Identify your bottleneck. Walk your process (or map it if remote). Where does work pile up? Where do people wait? Calculate the capacity of your top three suspects. The lowest number is your constraint — and your highest-leverage improvement opportunity.


Want to find bottlenecks faster? ProcessModel simulates your process with real variability, identifies constraints automatically, and shows you exactly where to invest. Stop guessing, start knowing.

Finding bottlenecks manually takes hours. ProcessModel identifies them automatically — and shows you exactly how much throughput you'll gain by fixing them.

Find Your Bottleneck