Managing Workflow and Bottlenecks

The idea: output is limited by the slowest step

Any process with several steps can only deliver as fast as its slowest step. That slowest step is the bottleneck. Everything before it piles up waiting; everything after it sits idle waiting for work. Improving a step that is not the bottleneck feels productive but usually changes nothing about how much gets finished.

This is useful because it tells you where to focus. Instead of asking everyone to work faster, you find the one constraint and deal with it first.

A simple worked example

Imagine a small business that makes and ships custom orders. Each order passes through three steps, handled by one person per step. In this hypothetical case:

Preparing an order: 20 orders per hour
Packing: 12 orders per hour
Labeling and dispatch: 30 orders per hour

The whole line can produce no more than 12 orders per hour, the rate of the slowest step. In an 8-hour day that is 12 x 8 = 96 orders, no matter how fast preparation or labeling is. The preparer finishes 160 orders’ worth of work in those 8 hours, but only 96 can be packed, so about 64 sit waiting at the end of the day. Speeding up labeling from 30 to 40 per hour would change nothing. Raising packing from 12 to 15 per hour would lift the line to 15 per hour, or 120 a day, until preparation (20 per hour) or another step becomes the new limit.

The figures are invented, but the pattern applies everywhere, including service work. A designer who can create ten concepts a week but needs client approvals that take two weeks has an approval bottleneck, not a creativity one.

Step 1: Map the workflow

You cannot find a bottleneck in a process you have not described. Write the steps in order, from the moment a customer commits to the moment they have what they bought. Include handoffs, approvals, waiting times, and any place where information or materials move between people or systems. A sheet of paper or whiteboard is enough.

For each step, note:

Who does it
How long it takes when working
How long work typically waits before it starts
What it needs (materials, information, a decision)

Often the waiting time between steps is larger than the working time inside them. A task that takes 15 minutes but waits three days for a sign-off is a problem of waiting, not effort.

Step 2: Look for the signs

Bottlenecks tend to leave visible clues.

Queues build up in front of one step, such as a growing pile of unpacked orders or unanswered emails.
Downstream people wait for work to arrive.
One person or machine is always busy while others have slack.
Deadlines slip at the same point repeatedly.
Customer complaints cluster around a particular kind of delay.
Overtime concentrates on one role.

Counting helps. Over a week or two, note how many items enter each step and leave it, and how long they wait. Numbers prevent the common mistake of blaming whatever feels busiest, which is not always what is limiting the system.

Step 3: Ask why it is slow

Once you know the step, find the cause. Typical reasons include:

Too little capacity: one person, one machine, a limited space.
Waiting for something: a decision, information, materials from a supplier, a delayed payment.
Rework: items are sent back because earlier steps left errors.
Interruptions: the person doing the step is constantly pulled away.
Unclear instructions: each time is a little different, so it takes longer than needed.
Batching: work is held until a full batch is ready, causing idle time.
Poor setup: tools or materials are hard to reach, requiring walking or searching.

The cause determines the fix. Adding people does not help if the problem is waiting for approvals.

Step 4: Ease the bottleneck

Try the cheapest, simplest options first.

Make sure the bottleneck is never idle. Keep it supplied with work and remove interruptions. Time lost at the constraint is lost for the whole system.
Remove non-essential work from it. If the packer also answers questions or fetches supplies, shift those jobs elsewhere.
Move work out. Another person, or an upstream step, may be able to do part of the task, such as prepping materials ahead of time.
Fix the quality before it arrives. Catching errors early prevents wasted effort at the constraint, which links closely to quality control.
Write the steps down. Clear standard procedures reduce hesitation and variation.
Rearrange the space or tools so less time is wasted moving or searching.
Simplify or standardize the offering. Fewer custom options can speed up the slowest step considerably.
Add capacity last. More hands, a faster tool or longer hours can work, but they cost money and add complexity, so be sure the step is truly the limit first. If you do add staff, consider whether delegation skills and a proper first hire are in place.

After each change, measure again. The bottleneck often moves to a new step, and that is a sign of progress, not failure.

Watch for common traps

Optimizing everywhere. Spreading effort across all steps usually delivers less than concentrating on the constraint.
Keeping people busy for its own sake. Non-bottleneck steps having spare time is normal and fine. Producing more than the bottleneck can handle just creates piles of half-finished work.
Ignoring waiting. Delays between steps are easy to overlook but often the biggest source of slowness.
Over-complicating the fix. A checklist or a rearranged table can beat an expensive tool.
Changing too much at once. If you change several things, you will not know which helped.

Customers and promises

Knowing your real capacity also improves what you promise. If your slowest step handles 96 orders a day, promising next-day delivery on 200 is a recipe for disappointment. Use your constraint to set realistic delivery times and communicate them clearly, as part of managing customer expectations. A shipping promise that you can actually keep is worth more than a faster one you often miss.

Keep a regular habit

Workflow is not fixed once. Seasons, new products, new staff and growth all shift where the limit is. Reviewing your process every month or quarter, perhaps as part of a regular business review, keeps small delays from becoming chronic. Improvement of this kind is gradual, and results vary by business, but finding and easing one constraint at a time is a reliable way to get more done with the same resources.