10 The Eight Wastes
“The goal is not to speed up the process, but to eliminate the time when nothing is happening.” — Shigeo Shingo
10.1 Why This Chapter Matters
A process can look productive while quietly burning time, money, and morale. People stay busy, but work still stalls, rework grows, and customers wait longer than expected. The Eight Wastes framework gives you a practical vocabulary for spotting these losses quickly. Naming a waste gives people a shared way to recognize and discuss it. Removing it still requires finding out why it occurs and testing a change.
10.2 Learning Objectives
By the end of this chapter, the student will be able to:
- Lean Six Sigma Principles and Tools Identify and distinguish the Eight Wastes in service and production processes Bloom:Understand
- Lean Six Sigma Principles and Tools Diagnose observed waste, supported inference, and unevidenced hypotheses using current-state map evidence, flow metrics, P90 tail behavior, and customer-promise performance Bloom:Analyze
- Lean Six Sigma Principles and Tools Frame an evidence-linked waste diagnosis that names competing explanations and the next observation needed to distinguish them Bloom:Analyze
10.3 Wash n’ Fold Project: Do
The Wash n’ Fold Baseline Performance Readout now shows five correction occurrences and a long release queue before Wash. Keep the completed readout and verified current-state map beside the diagnosis. They tell us where work stalls or circles back, but we still need to distinguish an observed loss from a suspected cause. The Eight Wastes give us names for those losses and help us decide what to investigate before designing a change.
In Chapter 8, we classified steps as value-added, required non-value-added, or non-value-added. Here we use the map and measurements to make a more specific diagnosis: what kind of waste appears, what evidence supports it, and what should change if a countermeasure works?
10.3.1 The Eight Wastes
The SPPM’s value-classification rule still applies: keep value-added work, make required non-value-added work more efficient, and investigate eliminating work that is neither valuable nor required.
What we want to identify now are eight different kinds of waste. Seven of these were introduced in (Ohno 1988). These are not the only possible wastes, but they are common enough to be worth memorizing. The acronym DOWNTIME helps:
- Defects
- Overproduction
- Waiting
- Non-Utilized Talent
- Transportation
- Inventory
- Motion
- Extra Processing
For each waste, we will start with an ordinary example and then return to Wash n’ Fold. Where a principle helps us distinguish the problem or choose a test, we will use it.
10.3.1.1 Defects
Definition 10.1
A defect is a failure of a product or service to fulfill a stated requirement or a reasonable expectation for use, including applicable safety requirements.
A pharmacy fills a prescription with the wrong quantity and must stop delivery, correct the package, and check it again. The correction uses time without producing another unit of customer value.
Unclear customer instructions, unstable work methods, inadequate training, worn equipment, and weak handoffs can all produce defects. Check the actual failure reason before choosing a response: clarify the order, improve the handoff, maintain equipment, or standardize a method that works. Count defects, affected units, and correction time separately, and check that the change improves first-pass yield.
Definition 10.2
First-pass yield, or FPY, is the proportion of units that complete a defined step or process without a defect, correction, scrap, or rework on their first pass:
\[ \mathrm{FPY}=\frac{\text{Units completed correctly on the first pass}}{\text{Units entering the defined boundary}} \]
State the process boundary, observation period, and denominator. One unit can contain more than one defect, so FPY is not the same as a defect count.
Sometimes quality checks can be required, for instance, when they are required by a contract or applicable regulation. But even so, they are still not adding value to the product since they would be unnecessary if the work were done correctly the first time.
At White Belt level, the key diagnostic question is simple: where could one bad tag or one bad handoff force the work to circle backward?
Five orders in the Wash n’ Fold log fail the packaging check and return for Fold correction; the map shows that return branch. The records establish rework, not whether bad tags, mixed items, or missed instructions caused it. Correction can increase lead time and reduce PCE, but the next observation should identify the failure reason before selecting a countermeasure.
10.3.1.2 Overproduction
Overproduction means making more than the customer needs now, or releasing more work than the next step can absorb.
A print shop runs 500 flyers for an event whose organizer ordered 200, or prints all of tomorrow’s jobs before the finishing station has room for them. The first case makes unwanted output; the second creates a pile of work in progress.
Pull authorizes production from actual demand and downstream capacity; keeping a machine busy is not sufficient authorization.
Large-batch incentives, forecasts treated as firm orders, and schedules that reward keeping a machine busy can release too much work. Compare release quantities with demand and downstream occupancy, then test smaller releases or a visible WIP limit. Do not replace a restrictive rule with unrestricted release: that may only move the queue.
The Wash n’ Fold case does not show finished laundry made for imaginary customers or too much work released downstream. Its serial rule holds the next wash batch until every order in the active batch clears drying, even when some machine positions are available. The large queue before Wash is waiting; it does not by itself prove overproduction. Inspect release times, available positions, and downstream occupancy before applying this label.
10.3.1.3 Waiting
Waiting is time when a unit of work is ready for a next step but cannot move forward. It occurs in manufacturing and service work alike.
A completed permit application sits in an inbox for two days before the next reviewer opens it. The applicant experiences those two days even if reviewing the form takes only minutes.
Batch release rules, uneven arrivals, missing information, and mismatched handoffs can make work wait. Measure the queue by step before changing a rule; then test smaller batches, clearer handoffs, or a controlled release signal. Check both the local wait and the customer’s end-to-end lead time so a reduction in one queue is not merely a transfer to another.
The Wash n’ Fold map measures a 501.7-minute mean release wait before Wash, compared with 3.6 minutes before Fold and package; other waits occur before Dry and notification. This is the clearest observed waste in the case. Chapter 9’s resource-workload estimate suggests enough physical capacity for average daily demand, yet a morning arrival burst and the serial release rule can still leave orders waiting. Observe queued orders and idle positions together, then ask what would happen downstream if release changed.
10.3.1.4 Non-Utilized Talent
Non-utilized talent occurs when people’s knowledge, skills, or judgment could improve the work but the organization fails to use them.
A clinic receptionist notices that patients repeatedly arrive without a required form and proposes a clearer reminder, but nobody with authority hears or tests the suggestion. The problem persists even though someone close to the work sees it.
Mapping the process requires learning how people actually do the work, not only how a procedure says they do it.
Narrow decision authority, a suggestion box with no follow-up, and work targets that leave no time to improve the method can suppress useful knowledge. Ask operators what they observe, test promising suggestions with them, and report what happened. Idle time establishes waiting; it does not by itself establish unused talent.
In Wash n’ Fold, ask whether operators have noticed the effects of the release rule and whether their suggestions are considered. The timing log cannot answer that question. Trace what happens to a concrete suggestion before diagnosing this waste.
10.3.1.5 Transportation
Transportation waste is unnecessary movement of materials or information between places.
A repair shop carries each finished part across the building for inspection, then carries it back to the same bench for packing. Neither trip changes the part in a way the customer values.
The named activities alone can hide a route that sends work back and forth across a room; the process map should make that route visible.
A distant layout, separated handoff points, and poorly placed storage can create extra trips. Trace the route and count trips before moving stations or changing a handoff; compare travel time and end-to-end performance afterward.
In Wash n’ Fold, laundry must travel from intake through sorting, washing, drying, folding, and staging. Excessive travel is possible, but the current records contain no station distances or trip counts. Observe the physical route before proposing a layout change or claiming a transportation saving.
10.3.1.6 Inventory
Inventory waste is inventory in excess of what the process needs, including unfinished work between steps. Some supplies and buffers are necessary; the question is whether the amount helps or obstructs delivery.
A repair desk accepts dozens of devices into an open pile while its technicians can complete only a few each day. The pile takes space, hides priority, and makes customers wait longer.
Queues and open-order counts show where unfinished work accumulates. A pull rule limits new releases to what downstream steps can handle.
Large releases, uncertain priorities, and bottlenecks can produce excess WIP. Track open work over time and its location, then test a visible limit or smaller release with a way to keep usable capacity supplied. Check lead time and completion as well as WIP: a smaller pile alone is not proof of better service.1
In Wash n’ Fold, the inventory of interest is unfinished orders awaiting release, partly processed loads, and orders awaiting notification, not just detergent on a shelf. The simulated Friday has 21 unfinished orders at close. That snapshot establishes open work, but the amount to reduce and the right place to limit it require a flow explanation. Finished orders awaiting pickup occupy space, although pickup lies outside the measured boundary.
10.3.1.7 Motion
Motion waste is unnecessary movement by the person doing the work, such as searching, reaching, or walking for a needed tool. Transportation concerns movement of the work or materials; motion concerns movement of the worker.
A laboratory technician walks to a distant cabinet for a label after preparing every sample because labels have no assigned place at the bench. Each sample triggers the same avoidable trip.
An observer may have to look inside a named step to see repeated searches and trips that the map alone conceals.
Poor workplace organization, misplaced supplies, and clutter often produce motion waste. Put frequently needed items at the point of use, give them clear locations, and measure retrieval time before and after. The next chapter develops 5S as one way to make that arrangement reliable.
The Wash n’ Fold setup method uses one shared supply jug that must be found, fetched, measured, and passed among washer positions. That context makes avoidable motion plausible, but the log does not separately time each search or trip. Observe setup to estimate the local saving, then check lead time and promise performance; PCE alone cannot establish which customer outcome improved.
10.3.1.8 Extra Processing
Extra processing is work beyond what is needed to meet the customer’s need and any legitimate process requirement.
A team copies the same address into three systems and prints a paper form that no one uses. The extra entries and printout do not improve the customer’s result.
Duplicate checks, unclear standards, and compensating for earlier defects can make extra work look responsible. People often justify an added step as being “careful.” Sometimes it is careful; sometimes it compensates for a weak upstream process. Ask what each step accomplishes before removing it; test a simpler method while monitoring defects, lead time, and customer outcomes. A shorter check is no improvement if it lets an incorrect order through.
At Wash n’ Fold, duplicate checking, re-tagging, repeated handling, and unnecessary sorting are possibilities to investigate, not recorded findings. Observe their purpose and frequency before calling them extra processing. The recorded Fold correction is already defect-related rework; do not count it again as a separate measured loss.
With all eight categories in view, ask two questions of the Wash n’ Fold map:
- Where would this waste appear on the Wash n’ Fold map?
- What local measure would test whether we reduced it, and what customer outcome would tell us whether that helped?
10.3.2 Wash n’ Fold Waste Summary
The current-state map tells us where to look first. The mean release wait before Wash is 501.7 minutes, compared with 3.6 before Fold and package. Among the 21 promise-eligible orders, median operating lead time is 424.5 minutes, P90 is 676.2, and six orders miss 5:00 p.m. The long tail matters because a reasonable-looking middle does not keep the promise for those six customers.
The table summarizes the first-pass diagnosis. An observed condition is a place to investigate, not automatic proof that the proposed cause or countermeasure is right. Some categories remain unestablished because this simulated Friday does not record the observations needed to assess them; there is no need to fill every box with a claim.
| Waste | Evidence status in this simulated Friday | First useful check |
|---|---|---|
| Defects | Observed: five packaging-check failures lead to Fold correction | Correction count and reason |
| Overproduction | Not established: the serial release rule holds work before Wash; a queue alone does not prove excessive upstream production | Release times, available positions, and downstream occupancy |
| Waiting | Observed: mean release wait 501.7 min; notification wait 37.9 min | These waits, then eligible-order P90 and promise violations |
| Non-Utilized Talent | Hypothesis: operator knowledge may be excluded from release decisions | Ask operators what they would change and why |
| Transportation | Not evidenced as excessive: movement between stations is necessary, but distances are not recorded | Trace the physical route before proposing a layout change |
| Inventory | Observed: 21 unfinished orders at close; whether and how to reduce it requires a flow explanation | Open-order counts over time, not just one snapshot |
| Motion | Setup context identifies a shared supply jug; avoidable searching is a hypothesis, not a timed finding | Observe fetching and measuring within setup |
| Extra Processing | Not evidenced separately: do not invent duplicate checks or count correction twice | Observe whether checks exceed customer or process requirements |
In my experience, people’s first thought about improvement is often to make a value-added step faster, perhaps by buying an expensive machine. At Wash n’ Fold, faster washers would be an attractive purchase, but they would not release orders that are waiting for permission to enter the machines already available. The map and baseline give us a better place to start investigating. This is the Least Cost Way: examine the waste before committing to equipment or staffing changes, while accounting for the cost, safety, and human consequences of any proposed change.
Using the table and your current SPPM, choose two wastes to investigate first. For each, identify the observed evidence, the cause that remains uncertain, a bounded next observation or low-cost test, and both a local measure and a customer-outcome measure. Explain why these two deserve attention before an equipment purchase. You may leave other categories unestablished.
Record your choices in your own notes or course workbook before opening the reference response.
Waiting before Wash is a strong first choice: its mean release delay is 501.7 minutes, and the serial rule is a plausible cause to test rather than a cause proved by the average alone. Observe release decisions, queued orders, and available washer and dryer positions; then test a controlled release change with a WIP limit. Track the release wait locally and eligible-order P90 and 5:00 p.m. promise violations for customers.
Defect-related correction is another defensible choice because five orders return from the packaging check. Record correction reasons before changing the handoff or work method, then track FPY, correction time, and operating lead time to ready-and-notified. Open WIP also deserves attention, but do not count the same held orders as an independent cause of their measured wait. The case does not yet establish excessive transportation, extra processing, unused talent, or overproduction.
10.3.3 Transfer Practice: Breakfast Service
The following breakfast-counter scenario is fictional.2 Read it as you would an observation of an unfamiliar process: identify what it establishes, keep uncertain causes open, and notice the safety concern.
A customer orders one slice of wholemeal toast, uncut, with butter. The cashier writes the order on a slip and places it in the preparation rack behind another ticket. Only “one toast, butter” is visible to the worker at the toaster. The worker selects one slice of white bread, the counter’s usual choice, and puts it in the toaster. While the toaster runs, the worker stands beside it. Clean trays for the next orders sit at a nearby station; whether preparing them during this interval would be safe and useful has not been tested.
When the toast is ready, the worker carries the plate to a second counter for butter. Three opened milk containers from earlier preparation crowd that counter, though no current order needs milk. The worker opens one drawer, then another, looking for a knife. The toast sits on the counter while they search for a knife to spread the butter. After finding it, they butter the slice and leave it uncut. On the way back to the customer, the worker steps over an appliance cord stretched across the walking route.
The customer points out that the bread is wrong and sends the plate back. The worker must make the order again while the customer waits.
Classify the wastes the story establishes, identify one category that still needs more evidence, and address the cord hazard. For each uncertain explanation, state what observation would help distinguish it. Write your diagnosis before reading the candidate revision below.
A candidate revision to inspect. The preparation rack holds each ticket with the full order visible. The worker reads the complete slip, selects one slice of wholemeal bread, and leaves it uncut. The cord is secured outside the walking route before service resumes. Butter and the tools actually needed for breakfast orders are placed where they can be reached without carrying a plate to the second counter; unneeded open containers are cleared from the work surface according to the food-safety method. While the toaster runs, the worker prepares the next clean tray only if that task can be done safely without missing the toast or confusing orders.
Compare this sequence with your diagnosis. Which evidenced trips, searches, errors, and extra steps does it address, and which could remain? Would the new placement create another safety or quality problem? Under comparable demand, track correct orders, elapsed time from order to delivery, and unnecessary trips before deciding whether the change helped. This is a proposed method, not a reported test result.
The wrong bread is a defect; the avoidable carry is transportation; searching is motion; and unnecessary open supplies are inventory. The finished toast waiting while the worker searches is a delay to the order. The obscured ticket supports investigating the handoff rather than assuming the worker ignored a clear instruction. The scene does not establish overproduction, extra processing, or non-utilized talent. To investigate those categories, observe preparation quantities, the purpose of added steps, and what happens to worker suggestions. The cord requires a safety response. The worker standing beside a running toaster does not by itself establish avoidable waiting; a second task must be safe and compatible with the first.
Choose one observed loss from breakfast service or a process you know. Record the observation, its waste classification, one supported explanation, one competing explanation, and the next evidence that would help distinguish them. Record your response in your own notes or course workbook.
10.4 Exercises
1. Bloom:Understand Define four of the Eight Wastes in your own words and give one realistic example of each from a service process.
Answers vary. A strong response uses accurate definitions and examples that match the waste category clearly (for example, waiting as idle queue time, defects as rework-triggering errors).
2. Bloom:Analyze In Wash n’ Fold, mean release wait is 501.7 minutes; five Fold corrections were recorded among 39 orders; six of 21 eligible orders miss the promise. Complete a Waste Diagnosis Brief: identify two evidenced wastes, one supported explanation, one competing explanation, and the next observation that would help distinguish them. Explain why these observations do not prove overproduction.
Waiting and defect-related correction are evidenced. The serial release rule is a supported explanation: it may leave usable capacity idle while orders wait before Wash. Insufficient effective capacity is a competing explanation. Observe release decisions, queued orders, and resource occupancy under comparable demand to distinguish them. A queue alone does not show that upstream production outruns downstream capacity, and the correction records do not establish incorrect tags or mixed orders.
3. Bloom:Analyze A team proposes buying faster equipment before mapping or measuring waste. Analyze this proposal using the least-cost-way principle.
The proposal is usually weak without prior waste analysis. If most lead time loss is from waiting, transport, or rework, faster equipment may add little value versus lower-cost flow and standardization fixes. Map and baseline first, then justify capital only if the bottleneck remains equipment-limited.
4. Bloom:Understand Define each of these waste pairs in one sentence each and state one practical difference:
- Waiting vs. Inventory,
- Motion vs. Transportation,
- Defects vs. Overprocessing.
- Waiting is idle time before the next needed action; Inventory is unfinished work or material held in the system. A queued claim can be both, but the terms describe its delay and its accumulated presence, respectively.
- Motion is unnecessary movement by a person within the work; Transportation is unnecessary movement of the work, customer, information, or material between places.
- Defects create incorrect or incomplete output that requires correction; Overprocessing adds work beyond what the customer or process requires even when the output is not defective.
5. Bloom:Analyze In a claims process, analysts repeatedly request missing fields from customers. Classify the primary waste and one likely secondary waste, then justify both choices.
The primary waste is Defects because incomplete submissions require correction. Waiting is a likely secondary waste while the claim sits for the missing information; repeated checking or duplicate entry could also create Overprocessing if the observation supports it. A strong answer separates what the scenario establishes from what would still need observation.
6. Bloom:Analyze A manager says, “If everyone stays busy, waste is low.” Analyze why this claim can be misleading in a queued system.
Local activity can remain high while work accumulates between steps, customers wait, and downstream capacity is overwhelmed. Releasing more work to keep one role busy may increase WIP and lead time without increasing completed output. Judge the system with flow, quality, and customer measures, not utilization alone.
7. Bloom:Understand Explain why waste classification must be tied to customer value and process boundary, not only local team effort.
An activity can look productive to one team while adding nothing the customer values or merely shifting delay outside that team’s boundary. Naming the customer, unit, start, and stop keeps the classification attached to the end-to-end result and prevents local effort from being mistaken for value.
8. Bloom:Apply For one service process you know, list two likely “invisible” wastes that customers feel but teams often normalize.
Answers vary. For each waste, name a specific customer-visible effect, an observable process sign, and the waste category. Examples include a request waiting unacknowledged in a queue, repeated transfer of the same information, or correction caused by an ambiguous intake field. Do not infer a cause or fill an unused waste category merely to complete the list.
9. Bloom:Analyze The Wash n’ Fold team could place supplies at each washer or change the serial release rule. Which observed or suspected waste would each change address, and which seems more likely to improve the 5:00 p.m. customer promise? Name the evidence you would check before choosing a test.
Point-of-use supplies would address setup-related motion and searching. Changing the serial release rule would address the much larger observed release wait and is therefore more likely to improve the customer promise. Before testing, check supply-retrieval time, order-release decisions, queue length, resource occupancy, demand, promise eligibility, and safety or workload effects. This evidence supports a bounded comparison; it does not prove the release rule is the sole cause of lateness.
10.4.1 Journal and Reflect
Write down one waste category that you think is easiest to overlook in your own environment and one sign you would look for to make it visible.
Which empty category tempted you to invent an example? Explain how you would defend “not evidenced” to someone who expects a completed box for every waste.
Identifying waste shows where the process loses time or effort; the Waste Diagnosis Brief preserves what is observed, what is inferred, and what must be checked next. Chapter 11 uses that brief to choose and specify a small operating design: how work moves, what authorizes it to start, and what workplace conditions will support the new method.
10.5 Chapter Summary
- The Eight Wastes are most useful as a way of reading a real process, not as an abstract memorization exercise.
- In Wash n’ Fold, release waiting and correction work are evidenced; the size of the release delay points toward a rule to investigate, not an automatic equipment purchase.
- Pair each proposed waste reduction with a local measure that tests its mechanism and a customer-outcome measure that checks whether it helped; use PCE and WIP to support that interpretation.
- The disciplined next move is not to buy speed everywhere; it is to identify the least-cost waste reductions that fit the losses the map actually shows.
Inventory can appear as an asset on a balance sheet, but a larger balance does not by itself mean better performance. Measures such as inventory turnover ask how quickly a business converts inventory into sales; a process change still needs to be judged by its actual costs and results.↩︎
Bruce Hamilton created Toast Kaizen, which inspired the use of toast making as a practice case here; see GBMP’s profile of Hamilton. The breakfast-counter scene and proposed revision are written for this chapter.↩︎