Step 2 — Analyze the Current Situation
How well you analyze the current situation determines how successful the kaizen will be. That means going to the actual place, observing the actual object, at the actual time — and then applying the right analytical tools to turn what you see into data you can act on.
Step 1 identified the opportunity. Step 2 is about understanding it deeply enough to improve it. That requires two things: the right posture when observing — the kaizen attitude — and the right tools for turning observation into structured analysis. Most kaizen failures are failures of Step 2: the observer relied on memory, skipped the gemba, or treated a qualitative impression as a quantitative fact. The tools in this section exist precisely to prevent that.
By the end of this section, you should understand:
- what the three actuals (genchi genbutsu) mean and why they matter;
- the three elements of analytic thinking — quantify, classify, detail — plus MECE;
- when and how to use Therblig motion analysis;
- how to fill in a Work-Element Analysis Sheet and what each column accomplishes;
- the seven-step procedure for conducting a time study on the plant floor;
- how standardized work functions as a kaizen lens;
- the scope of equipment-loss and material-flow analysis.
1The Kaizen Attitude — Three Actuals
Before any tool is picked up, the observer's posture determines the quality of everything that follows. The Toyota kaizen course calls this out explicitly, because experience is both an asset and a trap.
Genchi genbutsu — literally "actual place, actual object" — is the principle that facts must be gathered at the source. In the kaizen context it extends to three actuals: the actual place where the work happens, the actual object being worked on, and the actual time the problem is occurring. "Seeing is believing" is how the course states it.
Going to the actual place is not simply a matter of proximity. It means observing until you know — without having to look — what the operator will do next, where they will move, and what motion they will make. Second-hand reports, video from last week, or a manager's summary are not substitutes. Like the scene of a crime: if you see it first hand, you do not spend hours reconstructing what happened.
Go to the gemba. Observe at the workstation, machine, or point where the problem occurs — not in a conference room.
Check the actual part, tool, or material. Do not rely on other people's descriptions or substitute samples.
Observe when the problem is occurring. Reconstructing what happened after the fact introduces guesswork that cannot be eliminated.
Genchi genbutsu as three conditions for reliable observation. All three must be present for the analysis to rest on fact rather than inference.
What to notice: these are preconditions, not techniques. No analysis tool compensates for observing the wrong place, the wrong object, or a reconstruction of events rather than the event itself.
Three disciplines reinforce the three actuals. First: do not be swayed by preconceptions. The more experienced a supervisor, the stronger the pull to rely on experience — "we've always done it this way," "we tried that once." Particularly on topics with emotional charge, it is almost impossible to conduct accurate analysis without first setting aside the prior assumption and looking freshly. Second: observe thoroughly. Difficult problems have multiple causes, often complex and interacting. Experience alone will not find them; thorough observation is required. Third: maintain a calm attitude. Kaizen is not a competition. If you lead it as an emotional event, rational judgment deteriorates. Separate the people from the problem, and the feelings from the facts.
A useful self-test at the start of an observation: can you state — specifically and in numbers — what the current situation is? If the honest answer is "not yet," that is the point of genchi genbutsu. The observation period is not over until the answer is yes.
2Analytic Thinking — Quantify, Classify, Detail, MECE
Analytic thinking means stating the facts as they are — not overlooking anything, not pushing opinion into observation. The Toyota course identifies three elements. Modern practice adds a fourth that makes classification rigorous.
Quantify. Express things in numbers. "We have a lot of defects" is qualitative — and in kaizen, it is not enough. "We have 25 defects per 500 parts, 5% of output" is quantitative. The difference is not pedantic: you cannot set a target, measure progress, or confirm an improvement without a number. In daily manufacturing, qualitative language is the norm. In kaizen, it is always replaced.
Classify. Organize data into categories. Knowing the defect rate is 5% tells you little until you break that 5% into categories: dimensional accuracy, surface finish, surface class, cosmetics. Each category then points toward a different root cause and a different countermeasure. Without classification, the 5% figure is a single undifferentiated problem; with it, three or four distinct problems become visible.
Detail. Look at the level of granularity the problem requires. The familiar puzzle: a 3×3 grid looks like nine squares. Count all squares — 1×1, 2×2, 3×3 — and there are 14. Count all rectangles and there are 36. What looks obvious is more complex than it appears. In manufacturing, the operator who has run the same job for years looks smooth and easy. A detailed motion-level observation reveals why. Skill = good method + practice; the method is visible only through detailed analysis.
MECE (Mutually Exclusive, Collectively Exhaustive) is the modern overlay on classification. A set of categories is MECE when no item falls into two categories at once (mutually exclusive) and every possible item fits somewhere (collectively exhaustive). Categories that overlap produce double-counting; categories with gaps let problems hide. The classic manufacturing MECE structure for equipment losses is OEE's three buckets.
A machine with 648 units of theoretical per-shift capacity loses output to three MECE buckets: Availability (breakdowns 100, changeover 20), Performance (cycle-time losses 65, minor stops 30), and Quality (scrap 20, rework 13). Net actual output ~400 units — a 38% loss. The categories are mutually exclusive (no loss is double-counted) and collectively exhaustive (every unit of lost capacity lands in one bucket).
What to notice: classification turns "we lost 248 units" into three separate problem statements that each point toward a different countermeasure. A Pareto of the seven sub-losses would then show that, say, proximity-switch failures account for 70% of breakdown losses alone — which is where the investigation goes first.
3The Analysis Toolkit — Seven Techniques
Depending on the goal, different tools are used to analyze current methods. The original Toyota kaizen course identified six; equipment-loss analysis and material-flow analysis bring the set to seven. Every technique applies to either the operation (how a person works on a product) or the process (the flow of materials through the plant). The four operational tools receive full treatment here; the two support tools are summarized.
| No. | Technique | Focus | Primary purpose | Depth in this guide |
|---|---|---|---|---|
| 1 | Motion Analysis (Therbligs) | Operation | Break work into its most basic motions; identify and eliminate wasteful motion at the micro level. | Full treatment below |
| 2 | Work Element Analysis | Operation | Break work into named elements; question each for necessity, location, sequence, and method. 5W1H + ECRS. | Full treatment below |
| 3 | Time Study | Operation | Measure time for each work element; identify where time is lost and establish a baseline for improvement. | Full treatment below |
| 4 | Standardized Work | Operation | Use existing SW forms (cycle time, sequence, SWIP) as an X-ray of current method; deviation = kaizen opportunity. | Summary below; SW guide |
| 5 | Equipment Loss Analysis (OEE) | Process | Classify machine losses into Availability, Performance, and Quality buckets; identify largest loss category. | Compact reference below |
| 6 | Material Flow Analysis | Process | Map lead time vs. value-adding processing time; expose waiting, conveyance, and queue as improvement targets. | Compact reference below |
| 7 | Operation / Process Analysis | Both | Analyze the ratio of value-added work; distinguish man work, machine work, and combined work; find separation opportunities. | Named; see IE references |
The full toolkit for Step 2. The first three — motion analysis, work element analysis, and time study — are the most used and require practice to build skill. Standardized work is treated in its own guide. Equipment loss analysis and material flow analysis are process-level additions that complete the set.
What to notice: techniques 1–4 share a common set of questioning concepts — 5W1H and ECRS. Learning those concepts once gives you the questioning logic for all four tools.
4Motion Analysis and Therbligs
Motion analysis breaks a job into its most fundamental motions — the most detailed level of observation possible. The specific technique is Therblig analysis, developed by Frank Gilbreth in the early 1900s and consisting of 18 symbols that name every basic motion a hand can make. The name "Therblig" is Gilbreth backwards, nearly.
A Therblig is one of 18 symbols representing a basic hand motion — the smallest unit of work observable to the naked eye. Motion analysis assigns a Therblig to every motion in a sequence, making waste visible at the level of individual hand movements.
The classic teaching example: picking up a pencil from a box and placing it on a table. Seven Therbligs:
| Motion | Description | Therblig name | Value-added? |
|---|---|---|---|
| 1 | Reach out hand toward pencil | Transport Empty (TE) | No — empty hand movement |
| 2 | Close fingers around pencil | Grasp (G) | No — auxiliary |
| 3 | Lift pencil from box / surface | Disassemble (DA) | No — auxiliary |
| 4 | Carry pencil to destination | Transport Loaded (TL) | No — conveyance |
| 5 | Position pencil onto surface | Assemble (A) | Yes — the intended act |
| 6 | Open fingers, release pencil | Release Load (RL) | No — auxiliary |
| 7 | Return hand to start position | Transport Empty (TE) | No — empty hand movement |
Of seven motions, only one — Assemble — is value-added. The other six are necessary auxiliaries or pure waste depending on the layout. That ratio is typical.
Even the simplest act — picking up and placing a pencil — decomposes into seven motions, only one of which adds value. The two Transport Empty motions are the clearest waste candidates: if supply and destination were positioned closer, both would shorten or disappear.
What to notice: Transport Empty motions at both ends of the sequence are a common pattern. Reducing them — by improving layout, kitting parts closer, or re-sequencing — is typically the first motion-kaizen target. Grasp, Disassemble, and Release Load are auxiliary; Assemble is the only value-adding motion.
The full Therblig set of 18 covers the complete range of hand motion: Search, Find, Select, Grasp, Transport Loaded, Position, Assemble, Use, Disassemble, Inspect, Pre-position, Release Load, Transport Empty, Rest, Unavoidable Delay, Avoidable Delay, Plan, and Hold. In practice, the most frequently encountered on the factory floor are Transport Empty, Grasp, Transport Loaded, Assemble, Disassemble, Release Load, Position, and Inspect.
Motion analysis is not appropriate for every situation. Use it when:
- the job is highly repetitive — if the job is not repetitive, motion analysis yields no actionable pattern;
- cycle time is very short and many small tasks are performed;
- the goal is specifically to eliminate wasteful hand motions (reaching, grasping, positioning);
- launching a new program or changing to a new process requires establishing the best motion pattern from the start.
One limitation: Therblig analysis contains no time element. It reveals what motions occur and in what order, but not how long each takes. For that, a time study is also required.
The value-added motions in most manual assembly jobs are a small fraction of total motion. Typically only Assemble, Use, and sometimes Inspect directly add value. Everything else is auxiliary or waste. Marking which Therbligs in a sequence are value-added — and counting the ratio — often produces an immediate improvement idea without any further analysis.
5Work-Element Analysis and the Analysis Sheet
Work-element analysis is the most versatile of the four tools. Any job can be broken into work elements and analyzed. Where motion analysis works at the level of individual hand motions, work-element analysis works at the level of named actions — "obtain part," "carry part to machine," "load machine," "push start switch." Each element may comprise several Therblig motions.
A work element is a named unit of work — a discrete action in a job sequence. There is no fixed rule for how fine to make the elements; the right level of detail depends on the purpose. Finer elements expose more waste and small-second improvements. Coarser elements are better for writing work standards and teaching. Both extremes have uses; the key is choosing consciously.
A typical element breakdown for a punch / spot-weld / inspect / pack operation might run: obtain part A, carry part A to punch machine, obtain part B, punch part A, carry to welder, spot-weld, carry to inspection table, inspect, carry to pack chute, pack. A finer breakdown distinguishes each pick-up from each carry; a coarser one combines sub-steps. More elements force you to look more clearly at each small action and to identify the problems associated with it.
The working tool is the Work-Element Analysis Sheet. It carries the whole method on one page: work elements listed in sequence, then columns for problems, for 5W1H questioning, for ideas, and for ECRS action. The sheet does double work — it records the current state (Step 2) and begins to generate improvement ideas (Step 3). That is intentional: as you question each element, ideas will come. Record them immediately in the Ideas column and keep questioning; a more complete idea usually emerges once the whole job is worked through.
| No. | Work element | Problems | 5W1H — Question every element | Ideas | ECRS | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Safety | Distance | Dim. | Quality | Ease | Why? | What? | Where? | When? | Who? | How? | E | C | R | S | |||
| 1 | Walk to raw-material rack, obtain part A | 4 m | ✓ | Deliver parts to point-of-use at punch machine | ✓ | ||||||||||||
| 2 | Load part A into punch machine, push start | ±0.3 | ✓ | ✓ | Add location pin — eliminates realign step | ✓ | |||||||||||
| 3 | Carry punched part to spot-welder, load, weld | 2.5 m | ✓ | ✓ | Move welder adjacent to punch — combine steps 2–3 | ✓ | |||||||||||
| 4 | Carry to inspection table, inspect, pack into tote | 3 m | ✓ | ✓ | ✓ | Build-in check at weld fixture; inspect and pack at cell exit | ✓ | ✓ | |||||||||
Problems: safety · distance · dimension · quality · ease of operation. 5W1H: Why is this done? What is its necessity? Where is it best done? When? Who? How? ECRS: E eliminate · C combine · R rearrange · S simplify.
Four work elements across the row: name the element, flag problems by category, question it with 5W1H, record any idea, then mark the ECRS action. The first two elements are fully worked; elements 3–4 show further examples. This is the actual artifact used on the plant floor.
What to notice: the Distance column in the Problems section is doing heavy lifting here — four separate walks totaling nearly 10 meters for a single cycle. Walk distance is one of the most common wastes visible immediately on a completed sheet. The Ideas column is filled in while questioning, not after; that is how the sheet generates Step 3 material at the same time it captures Step 2 facts.
The five problem categories — safety, distance, dimension, quality, ease of operation — give each element a structured lens. Safety flags anything that creates risk. Distance records physical distances walked or reached. Dimension notes tight tolerances or gauge requirements. Quality marks elements where defects are produced or detected. Ease of operation flags ergonomic difficulty, awkward posture, or excessive force.
The 5W1H questions are applied to every element in turn. Why? — Is this element necessary at all? What? — What is its specific necessity? Where? — Is this the best place to do it? When? — Is this the best point in the sequence? Who? — Is this the right person to do it? How? — Is this the best method? Any question answered with "I'm not sure" is an improvement opportunity.
ECRS is the action decision: can this element be Eliminated entirely (most powerful), Combined with an adjacent element, Rearranged in sequence, or Simplified in method? Eliminate always comes first in that priority order. An element that survives elimination questioning is then a candidate for the other three.
Writing elements from memory rather than from direct observation. The sheet is an observation tool, not a reconstruction tool. Fill it in at the gemba, watching the operation run, not afterward at a desk. Memory systematically omits the small elements — the adjust, the check, the reach — that are often the richest improvement targets.
6Time Study
Time acts like a shadow: it reflects how the work is done. A change in method changes the time. That is why time is an unbiased judge of improvement — and why time measurement is a fundamental part of Step 2.
The most practical method for plant-floor time study is the consecutive method: start the stopwatch once and read the cumulative time at each measurement point, writing down the running total at the end of each element. Individual element times are calculated afterward as the difference between consecutive readings. This avoids the error of stopping and restarting the watch, which introduces handling time into the data.
Measuring points are the agreed start and stop points for each element. Every time study requires clear measuring points defined before measurement begins. Without them, different observers measure slightly different things, and the results cannot be compared. A measuring point is typically the moment a motion ends cleanly — the hand releasing the part, the button pressed, the part placed in the container.
For the representative time, Toyota's practice is to take the most-repeated time for each element — the time that appears most frequently across cycles, not the average and not the shortest. Note: some lean companies take the shortest non-wasteful cycle time as the target (representing the instance when everything went right without waste). Both approaches are valid in context. The most-repeated time is more conservative and more stable for unstable processes; the shortest cycle is the theoretical ideal. Know which convention your organization uses.
The seven-step procedure for conducting a time study on the plant floor:
More measurements do not mean greater accuracy, especially for unstable processes. On an unstable process, ten measurements capture ten different states; they do not average out to a reliable time. Stability first, then timing. If cycle time varies widely, the variation itself is the problem to address — before a representative time means anything.
7Standardized Work as a Kaizen Lens
When a job has already been standardized, the Standardized Work forms become powerful kaizen inputs. The Standardized Work Chart shows the work sequence and layout; the Standardized Work Combination Table shows manual time, machine time, and walk time for each element side by side; the Process Capacity Sheet shows each machine's theoretical and actual output per shift. Together they create an X-ray of the current method.
Looking at the work sequence reveals walking distance and return time. The Combination Table shows wait time — the operator standing idle while a machine runs — and the ratio of manual to machine work. Standard Work in Process (SWIP) shows where inventory accumulates in a cell. Any deviation from the standard is a fact, and a fact worth investigating: it either means the standard has a flaw, or something in the method is preventing the standard from being followed. Both are kaizen opportunities.
Where no standardized work yet exists, Step 2 analysis (time study, work element analysis) generates the data needed to create it. In this way, kaizen and standardization reinforce each other: there is no kaizen without a standard to compare against, and there is no standard without the measurement work of Step 2.
See the Standardized Work guide for a full treatment of the Process Capacity Sheet, Standardized Work Combination Table, and Standardized Work Chart — and for the principle that standardized work is the basis for all kaizen.
8Compact Reference — Equipment Loss and Material Flow
Two additional techniques extend Step 2 beyond the operational level to the process level.
Equipment Loss Analysis (OEE framework). When a machine is the constraint or the source of the problem, classify its losses using the three MECE buckets illustrated in Figure 5.2 above. Availability losses are unplanned downtime (breakdowns) and planned downtime within the production window (changeovers, planned maintenance). Performance losses are slow cycles and minor stops — the machine is running, but not at standard speed or without interruption. Quality losses are defects and rework produced. OEE = Availability Rate × Performance Rate × Quality Rate. A Pareto within the dominant loss bucket then identifies the specific problem to countermeasure — typically the top one or two items account for 60–80% of total loss.
Material Flow Analysis. Map the end-to-end lead time for a part or product family from raw material release to shipment. Then measure the actual value-adding processing time within that lead time. In most factories, processing time is 1–5% of lead time; the rest is queue, wait, conveyance, and inspection. The ratio of lead time to processing time is the first target: large ratios indicate systemic flow problems — lot sizing, push scheduling, poor layout — that cannot be resolved by improving a single workstation. This is the analytical foundation for value-stream mapping.
9Observing on the Plant Floor
The Session 3 practicum structure is the proven framework for running a plant-floor observation. Form teams of three to four; assign one theme per team and limit the range of observation to one or two employees. Target: first 30 minutes to identify and name all work elements; first hour to complete the time study of those elements; 90 minutes to complete motion analysis; two hours total.
Key analysis points while observing:
- Work element analysis: capture the degree of fluctuation in each element; note walking distance explicitly; distinguish time with work in hand from time with an empty hand; identify any abnormal work (work that occurs irregularly, outside the standard cycle); find the bottleneck element.
- Motion analysis: identify value-added motions, auxiliary motions, and non-value-added motions separately; observe both hands — the non-dominant hand is often idle when it could be contributing; note the direction of motion (reaching across the body is a common waste source); observe how materials are gripped and how the workpiece is adjusted before or after processing.
Five precautions for plant-floor observation: do not hinder the work in progress; maintain all safety requirements (PPE where required); complete the observation within the allotted time; ask the team leader about unclear points rather than interrupting the operator; do not disturb employees while they are working.
At the conclusion of observation, prepare four materials for the improvement discussion: the Work-Element Analysis Sheet (completed), the Motion Analysis Sheet (completed), the Time Observation Form (completed with most-repeated times), and a sketch of the work area showing the layout, the number of operators, process flow, and inventory locations. This package becomes the factual basis for Steps 3 and 4.
The goal of plant-floor observation is to see things that have never made an impression before — not to confirm what you already know. The experienced supervisor's challenge is precisely this: the job looks normal because it always has. Therblig notation and the Work-Element Analysis Sheet work partly because they force you to write down things that familiarity has made invisible.
Section summary
Step 2 begins with the kaizen attitude: go to the actual place, observe the actual object, at the actual time. Do not be swayed by preconceptions; observe thoroughly; maintain a calm attitude. Analytic thinking means quantifying (replace qualitative impressions with numbers), classifying (organize data into MECE categories — mutually exclusive and collectively exhaustive), and detailing (look at the level of granularity the problem demands). The equipment-loss waterfall is the model MECE classification: Availability, Performance, and Quality buckets that together account for every unit of lost capacity.
The analysis toolkit has seven techniques. The four core operational tools: Therblig motion analysis — 18 symbols naming every basic hand motion, exposing the ratio of value-added to non-value-added motion at the micro level; Work-element analysis with the Work-Element Analysis Sheet — name each element, flag it for problems (safety / distance / dimension / quality / ease), question it with 5W1H, record ideas, mark an ECRS action; time study — consecutive method, agreed measuring points, most-repeated time, seven-step plant-floor procedure; and standardized work as a kaizen lens — deviation from standard is a fact and an improvement opportunity. Two process-level additions: equipment loss analysis using OEE's three buckets and Pareto within the dominant bucket; material flow analysis exposing the ratio of lead time to actual processing time. The plant-floor observation structure — teams, 30-minute element identification, 60-minute time study, 90-minute motion analysis — produces the four-document factual package that Steps 3 and 4 run on.