Orders are going out late, and the first thing that comes to mind is that you are short of people. You hire someone, you pay their salary, and three months later the orders are just as late.
In any repeating path, the speed of the whole job is set by its slowest step. That step is called the bottleneck. Until you open up that one step, the rest can get as fast as they like and the output will not change.
How is a process bottleneck different from a business bottleneck?
A process is a repeating path that takes an input and turns it into an output, like the path from placing an order to delivering it, or the path from a lead to a customer. A business bottleneck stands one step further back and asks which part of the business has fallen behind the rest. A process bottleneck means zooming in on that business bottleneck and walking one specific path step by step to see where the problem is.
That difference matters. If you find out your operations are slow and stop there, your fix becomes pouring people and money into all of operations. But when you open up the path itself, you usually find that out of ten steps only one is stuck and the other nine are fine. Fixing that one is far cheaper and faster than treating the whole department.
Why the bottleneck is not the busiest spot
You would think the bottleneck is where people are working themselves to the bone, but it is usually the opposite. If you measure the life of a job from the moment it enters the path to the moment you hand it over, only a small part of that life was actual work and the rest of it sat in a queue. In a lot of teams that ratio is between 15 and 25 percent, which means three to five hours of waiting for every hour of real work.
So do not look for the busiest person. Look for the place where work has piled up in front of it and everyone behind it is waiting. That hill of half-finished work tells you where the bottleneck is more accurately than any report.
Is your bottleneck capacity or policy?
There are two kinds, and the treatment for each is different. A capacity bottleneck means a step genuinely cannot handle more, like one person who cannot get through more than a certain number of jobs a day. Adding capacity is the right answer there. A policy bottleneck means a rule you set yourself has slowed the work down, without anything actually running short.
In the Theory of Constraints, one of the oldest frameworks for finding bottlenecks, the most common constraint is that second kind. What we call a shortage of resources is usually a decision of our own that we have forgotten the reason for. Three of them come up again and again:
- Everything has to go through you. Even the smallest decision waits on one person.
- Work is collected and done once a day. That alone creates a queue that has nothing to do with the volume of work.
- One person holds the key. Only one person has the password, the access or the information, and everyone else waits behind them.
None of these is a shortage of people. Each is a rule that opens the queue the moment you lift it, without spending a rial.
An example from a small print shop
Picture a small digital print shop taking thirty to forty orders a day. Your quality is good, but you keep hearing that orders were ready late. Your first instinct is that you are one person short at the machines.
One day you draw the path of a single order on paper, from the moment it is placed to the moment it is handed over. The customer’s file is print-ready a few minutes after the order comes in, and the printing itself takes a few minutes more. But every order is stuck in one place, behind a rule you set yourself, that final approval of the artwork is yours and you review them all together in the evening. An order placed at nine in the morning sits until the end of the day waiting for one approval from you.
The bottleneck here is that evening-approval rule, not the machines and not the people. And the treatment is not hiring. It is enough to let one other person approve as well, or to do it three times a day instead of once at night. An order that used to wait ten hours now waits three at most, without a rial of extra cost.
If you had hired someone for the machines instead, the printing would have got faster and the orders would still have sat on your desk until the end of the day.
How to find your process bottleneck
These five steps work on any path, from delivering orders to producing content to answering customers.
- Pick one specific path and write down its steps. Not the whole business, just one path, say from order placed to order delivered. Put every step on paper in order so the whole chain is in front of you.
- Follow one real example through it. Take a single order and write two numbers at each step, how many minutes of work it actually got and how many hours it spent in the queue.
- The step with work piled up in front of it is the bottleneck. Where half-finished work stacks up and everyone right after it is waiting, that is the slowest link.
- Ask whether the slowness is capacity or policy. Does this step genuinely lack the capacity, or has a rule of your own slowed it down? The answer to that question changes the treatment completely.
- Lift the rule first, then add capacity if you still need it. Lifting a rule is usually free. And once you do, the bottleneck moves to the next point, so this is not a one-time job and you have to repeat it every so often.
Every repeating path has a slowest step that sets the speed of the whole output. The root of that slowness is usually a rule you set yourself that keeps work sitting in a queue for no reason. So the next time you are about to add a person to a slow job, first ask whether capacity is genuinely short here or a rule of your own has slowed the process down. Until you answer that, you could pay salaries for months without the output getting any better.