Batch and queue
One-piece flow
One-piece flow, also called single-piece flow, means making and moving one unit at a time through a connected set of steps. Instead of finishing a whole batch at one station before passing it on, each item goes to the next step as soon as that step can take it. The focus is the product's path through the whole process, not the output of one machine or department.
This guide covers how one-piece flow works, a worked comparison with batches, takt and cycle time, the benefits to check, production and transfer batches, how flow relates to pull, where it fits, the conditions it needs, a six-step pilot, how to measure it and common misconceptions. All numbers are illustrative.
Key takeaways
One-piece flow changes when the next step can start, not how long each step takes.
The gains come from less waiting, less work in process and faster feedback on defects.
It needs steady work content matched to takt, short changeovers, reliable equipment and standard work.
Where steps cannot be joined directly, a FIFO lane with a fixed limit keeps the flow controlled.
Pilot on one product family and judge the whole result, not one station's output.
What is one-piece flow?
Four traits define it:
Single transfers
Each finished unit moves on without waiting for a transfer batch.
Linked steps
Operations are arranged and paced to remove delays between them.
Low work in process
Unfinished items cannot pile up freely between operations.
Demand-based pace
The process makes what the next step needs, not what spare capacity allows.
Continuous flow is a close relative. Lean writers sometimes use it more broadly to include small, steady batches with little delay. Strict one-piece flow moves single units; real lines often mix single transfers with controlled buffers or a batch step that cannot be avoided. The useful question is whether work moves steadily or spends time waiting.
Batch vs one-piece flow: a worked example
Three stations assemble, test and package a product. Each step takes 1 minute per unit. The order is 10 units, each station has its own operator, and setup, travel and defects are ignored:
Transfer batches of 10 · order done at 30 min
One unit at a time · order done at 12 min
Result
Batches of 10
One at a time
First unit reaches testing
After 10 minutes
After 1 minute
First finished unit
After 21 minutes
After 3 minutes
Whole order finished
After 30 minutes
After 12 minutes
Total processing work
30 station-minutes
30 station-minutes
The work content is identical. What changes is when downstream work can start: with single transfers the three stations overlap. In general, with n units and k equal steps of t minutes, batches finish after n × k × t and single-unit flow after (n + k − 1) × t. Real lines add setups, travel and variation, so treat this as the shape of the effect, not a forecast.
Takt time and cycle time
Flow needs a pace to work to. Takt time sets it from demand:
takt time = available production time ÷ customer demand
Cycle time is different: it is the measured time a step or process takes to complete one unit. A connected cell works when each step's cycle time fits inside takt. If one step runs longer, it becomes the pace of the whole cell, and line balancing shows how to spread the work. The cycle time guide covers measuring it.
Benefits to check
Each benefit has a mechanism and a way to confirm it:
| Benefit | Where it comes from | How to check it |
|---|---|---|
| Shorter lead time | A finished part no longer waits for the rest of its batch | Record when each unit was ready for the next step and when that step began |
| Less WIP and floor space | Queues between departments shrink | Count real units at each handoff, not only system records |
| Faster quality feedback | The next step sees a defect within minutes | Track the time from making a defect to finding it |
| Visible bottlenecks | With no piles to hide behind, a slow or unreliable step shows at once | Log each stop with its time, cause and response |
| Flexibility within a product family | A cell can switch variants if methods and parts supply allow | Test the hard combinations, not only the easiest product |
| Better use of people | Work is shared across the cell instead of keeping each machine busy | Review walking, reaching and support work with the operators |
| A baseline for improvement | Standard work sets sequence, takt and standard WIP | Keep a change log of problems, changes and results |
Shorter lead time
A finished part no longer waits for the rest of its batch
How to check it: Record when each unit was ready for the next step and when that step began
Less WIP and floor space
Queues between departments shrink
How to check it: Count real units at each handoff, not only system records
Faster quality feedback
The next step sees a defect within minutes
How to check it: Track the time from making a defect to finding it
Visible bottlenecks
With no piles to hide behind, a slow or unreliable step shows at once
How to check it: Log each stop with its time, cause and response
Flexibility within a product family
A cell can switch variants if methods and parts supply allow
How to check it: Test the hard combinations, not only the easiest product
Better use of people
Work is shared across the cell instead of keeping each machine busy
How to check it: Review walking, reaching and support work with the operators
A baseline for improvement
Standard work sets sequence, takt and standard WIP
How to check it: Keep a change log of problems, changes and results
Do not book every removed rack as a saving. Freed space and time are only cash savings when a real cost goes away; otherwise record what the space and time are now used for.
Production batches and transfer batches
These are two separate decisions:
Production batch
How many units are made before switching a process to another product.
Transfer batch
How many units move to the next step together.
Moving single units does not mean changing products after every unit. A cell can run the same model all morning and still pass each unit on alone. Write both numbers down at each step; often the first test is changing the transfer rule while the production schedule stays the same.
Flow and pull
Flow is about how work moves through steps. Pull is about the signal that tells an upstream step to make or refill what a downstream step has used. A line moving single units still needs a pull signal, or it will make too much.
Toyota's just-in-time system uses small, set stocks that downstream steps draw from and upstream steps refill, often with kanban cards. That corrects the idea that flow means every stock point must be empty. The design question is which stock has a clear purpose and how its refill is controlled.
Batch and queue vs one-piece flow
The patterns at a glance:
| Point | Batch and queue | One-piece flow |
|---|---|---|
| Handoff | Several units move together | Single units move on |
| Between steps | Queues build up | Links limit waiting |
| Coordination | Departments work apart | Consecutive steps work together |
| Improvement focus | Each step's own speed | The product's movement across steps |
| Problems | Hidden by inventory | Visible at once |
Handoff
Several units move together
One-piece flow: Single units move on
Between steps
Queues build up
One-piece flow: Links limit waiting
Coordination
Departments work apart
One-piece flow: Consecutive steps work together
Improvement focus
Each step's own speed
One-piece flow: The product's movement across steps
Problems
Hidden by inventory
One-piece flow: Visible at once
Not every batch process is badly run. Compare the real route, transfer sizes, setup needs and demand, and look at the full result rather than the measure that flatters one option.
Where one-piece flow fits
Illustrative candidates. For each, ask whether the next step can take a single item and what stops it today:
Assembly then test
Pass each assembly to a ready test station, checking test time, failure handling and equipment availability.
Test then packaging
Pack accepted units one by one instead of letting a pile build, checking labeling and order grouping.
Product-family cell
Group the steps for related products, checking the route and work content of every variant included.
Office requests
Complete and hand on single requests instead of holding them for a daily bundle, keeping required approvals.
Mixed route
Single transfers in assembly with a batch step such as heat treatment upstream, joined by a controlled interface.
The last case is common. Some processes, such as ovens, plating or painting, run in batches for good reasons. Improve flow where you can and manage the remaining joints on purpose. To pick a candidate, follow one real item through its whole route and note where it stops, who decides when it moves and what information travels with it.
Conditions flow depends on
Removing inventory removes the cushion that hid these problems, so put them in place first:
Work matched to takt
Time the full pattern, including refills, variant differences and interruptions, not only the easiest cycle.
Short changeovers
SMED moves setup tasks that can be done while the machine runs out of the stopped time.
Reliable equipment
TPM brings operators and maintenance together so connected machines are ready when needed.
Clear standard work
The sequence, takt and standard WIP are written down with the people who do the work.
FIFO lanes where steps cannot connect
When two steps cannot sit side by side, a FIFO lane keeps the sequence. The first item in is the first out, and a full lane tells the supplying step to stop. For each lane, write down:
Max 5 · first in, first out · full lane stops step A
- Purpose
- Why the two steps cannot be joined directly.
- Contents
- Which items may sit in the lane.
- Limit
- The maximum number of units, marked on the lane.
- When full
- The supplying step stops and the reason is investigated.
- When empty
- The receiving step signals for help.
- Owner
- Who checks that the lane still serves its purpose.
A lane is a visible design choice the team can challenge, not permission to store material wherever space is free.
Piloting one-piece flow in six steps
Start with a bounded pilot before rearranging a whole plant:
- 1
Choose a product family. List the products included, where the pilot starts and ends and the customer need it must meet. Note products left out because their routes differ.
- 2
Walk the current route. Follow sample units through processing, waiting, moving, inspection and rework, and write down the real handoff rules, including informal ones. A value stream map is a good format.
- 3
Set a baseline. Agree measure boundaries and record lead time, WIP, good output, quality losses and stops, across shifts and variants.
- 4
Design the connection. Describe the work sequence, transfer rule, refill signal and any FIFO lanes. Review layout, handling and equipment with the people who do the work.
- 5
Prepare the response to problems. Name who supports the cell, how help is called and how suspect material is held, before removing the inventory that used to separate the steps.
- 6
Trial and review. Run under controlled conditions, compare with the baseline, record what failed as well as what improved and revise before expanding.
Set the decision rule before the pilot starts, for example: shorter lead time with no drop in good output, quality or operator workload. Otherwise an attractive layout becomes the only proof of success.
Measuring the result
Use a small, balanced set of measures and read them together:
| Measure | What to set | Watch for |
|---|---|---|
| Lead time | Clear start and finish points | A shorter time inside the cell while the queue moves elsewhere |
| WIP | Units present at set locations | Stock moved off the floor but not gone |
| Good output | Good units completed against demand | Output rising at one station only |
| First-pass yield | Units passing without repair | Defects caught later instead of earlier |
| Stops | Time, cause and recovery | Problems hidden to keep the line moving |
| Labor effort | Total effort, including support work | Effort shifted onto material handlers |
| Operator feedback | Access, handling and workload | Steady output that depends on rushing |
Lead time
Clear start and finish points
Watch for: A shorter time inside the cell while the queue moves elsewhere
WIP
Units present at set locations
Watch for: Stock moved off the floor but not gone
Good output
Good units completed against demand
Watch for: Output rising at one station only
First-pass yield
Units passing without repair
Watch for: Defects caught later instead of earlier
Stops
Time, cause and recovery
Watch for: Problems hidden to keep the line moving
Labor effort
Total effort, including support work
Watch for: Effort shifted onto material handlers
Operator feedback
Access, handling and workload
Watch for: Steady output that depends on rushing
flow efficiency = value-adding time ÷ total lead time
Set the process boundary first. A required inspection may be needed work even though it does not change the product. The flow efficiency calculator does the arithmetic.
Common one-piece flow misconceptions
- 1
Flow means never stopping. Stopping to protect quality is part of the design. Passing on bad work is not flow.
- 2
Low inventory means no inventory. Standard work can include set in-process stock, and pull systems use controlled refill stock.
- 3
A compact cell proves flow. Check whether work really moves steadily and meets demand. Layout supports the method; it does not replace it.
- 4
Every step must flow one by one. Some batch steps stay. Manage the joint instead of forcing single units everywhere.
- 5
Flow is only for factories. The same idea applies to requests, claims and code changes that wait in batches.
Start with a bounded process and a measurable problem, and test whether the new method cuts waiting while keeping quality, output and workload steady.
Frequently asked questions
What is one-piece flow?
Why is one-piece flow faster than batches?
What is the difference between one-piece flow and continuous flow?
Does one-piece flow mean zero inventory?
Choosing Time Study Software?
Read the buying guide: what to look for, mistakes to avoid and questions to ask vendors.
Related guides
- Line balancing: how to split work so a line meets taktLine balancing explained: takt and cycle time, balance efficiency, the data to collect, yamazumi charts and assignment rules, a worked four-station example, workstation checks, software questions and a pilot plan.Read the guide
- Lead time: how to define it, calculate it and shorten itLead time explained: where the clock starts and stops, the common types, how to calculate it, ARO quotes, reorder points, the metrics worth tracking and how to shorten it without cutting quality.Read the guide
- Cycle time: definition, formula and how to reduce itCycle time explained: start and end points, the formula, cycle time vs lead time and takt time, percentiles, what makes it long and how to bring it down.Read the guide
- OKRs: how to write, run and score objectives and key resultsOKRs explained: what an objective and key result are, the quarterly cycle, OKR vs KPI, writing testable key results, illustrative examples for marketing, product and support, scoring, committed and stretch goals, a first-cycle plan and mistakes.Read the guide
Sources
Books and papers show their publication year. Web pages show the month we last checked them.
- 1.Mike Rother and Rick Harris, Creating Continuous Flow: An Action Guide for Managers, Engineers and Production Associates (Lean Enterprise Institute, 2001)
- 2.James P. Womack and Daniel T. Jones, Lean Thinking: Banish Waste and Create Wealth in Your Corporation (Simon & Schuster, 1996)
- 3.Jeffrey K. Liker, The Toyota Way: 14 Management Principles from the World's Greatest Manufacturer (McGraw-Hill, 2004)
- 4.Shigeo Shingo, A Revolution in Manufacturing: The SMED System (Productivity Press, 1985)
- 5.Seiichi Nakajima, Introduction to TPM: Total Productive Maintenance (Productivity Press, 1988)