Step 4 — Make the Plan
There are many ways to reach any goal. The job at Step 4 is to draw up several proposals, weigh them against quality, cost, delivery, and safety, choose the best, then translate it into a concrete action plan before anyone touches the process.
Steps 1 through 3 produced a root cause and a set of ideas. Step 4 turns those ideas into a plan. The distinction matters: an idea is a direction you might go; a proposal is a specific approach you can evaluate; a plan is the sequence of actions, responsibilities, and deadlines that puts the proposal into motion. Rushing from idea to implementation without this step is one of the most common reasons kaizen fails to hold.
By the end of this section, you should understand:
- why generating multiple proposals — not one — is a discipline, not a luxury,
- the distinction between work kaizen and equipment kaizen, and which takes precedence,
- the six directions of kaizen and what each one means in practice,
- how to structure an implementation plan with the four essential columns.
1Draw up several proposals
When you have identified the root cause of a problem, the temptation is to implement the first solution that looks workable. Resist it. There are almost always multiple ways to achieve the same goal, and the first one that comes to mind is rarely the best.
Consider a simple question: how many different ways can you join two steel sheets to each other? You might bond them with adhesive, caulk them, fasten them with nuts and bolts, weld them, or bind them with cord. Five approaches, each with different cost, strength, reversibility, and production impact. The point is not that five is the right number — it is that forcing yourself to generate options before evaluating any of them exposes the space of solutions and prevents premature commitment.
For each proposal, consider its effect on the four target areas: quality, cost, delivery (production volume and lead time), and safety. A proposal that reduces defects but doubles cycle time may not be a net improvement. Study the proposals as a set, then pick the one best aligned with the improvement objective.
In practice, "several" means at least three. Fewer and you haven't really explored the space; more than six and evaluation slows without adding clarity. The goal is a genuine comparison, not an exhaustive search.
There is also an element of trial and error once you get to implementation. Some proposals that look good on paper will not hold up in the actual work environment. Generating options at Step 4 means you have alternatives ready when the first attempt needs adjustment — and it usually does.
2Work kaizen vs equipment kaizen
All kaizen proposals fall into one of two categories: work kaizen or equipment kaizen. Knowing the difference — and the priority between them — shapes how you develop proposals.
Work kaizen improves the method: how work is distributed, sequenced, and performed. It is based on Standardized Work. It involves little additional expenditure. Anyone — team members, group leaders, engineers — can and should engage in it.
Equipment kaizen introduces or modifies machinery and automation. It requires a special crew or engineering team and involves capital expenditure. It is the right tool when work kaizen has reached its limit — not before.
Work kaizen is central to the development of the Toyota Production System. It determines the rules for redistributing and combining work, fixes the layout and storage location of parts with workability in mind, and improves operations directly. Because it requires no capital and can be executed by the people doing the work, it is faster, cheaper, and far more flexible than equipment kaizen. Main emphasis should always be on work kaizen.
Equipment kaizen — machinery changes, automation, new tooling — has its place, but only after work kaizen has been exhausted. Automating a bad method produces a faster bad method. The discipline is: try work kaizen first. If the constraint is genuinely in the equipment and not the method, then pursue equipment kaizen.
Reaching for equipment kaizen because it feels more decisive. Purchasing a machine or adding automation is visible and measurable in ways that method improvements sometimes are not. But it is also irreversible and expensive. Work kaizen should be exhausted before any capital solution is proposed.
3The six directions of kaizen
When evaluating proposals — whether work kaizen or equipment kaizen — there are six directions in which kaizen can move. Think of these as the fundamental moves available to you. They run roughly in priority order: elimination is always preferred over combination, combination over rearrangement, and so on. Try the earlier directions before accepting that a later one is necessary.
If a step, motion, or activity is judged unnecessary, remove it entirely. No improvement is more powerful — or more permanent — than eliminating waste rather than optimizing it.
Separate tasks that belong together, or merge tasks that are currently fragmented. Separate or combine — both are tools for getting to a simpler, more balanced method.
Change the sequence, layout, or form of work to make it easier to perform. Reduce walking distance, reposition materials, redesign tooling. Use the full capability of people and equipment.
Reduce variation in method, speed, tooling, and procedure. In TPS there is no kaizen without standardization — a method improved but not standardized will drift back.
Align the timing and flow of related processes so the right work arrives at the right place at the right time — the Just-in-Time principle applied at the method level.
Design the process so abnormalities are detected and stopped at the source — in-station process control / jidoka. Stop on defect; don't pass a bad part downstream.
The six moves available when developing a kaizen proposal, shown in priority order. Each card names the direction and its core logic.
What to notice: the numbering reflects priority. Elimination is always preferred — optimizing a step you could remove is waste on waste. Standardization appears at position 04 because any improvement at 01–03 must be stabilized before it can be the baseline for further kaizen.
4Each direction in practice
Eliminate. If the reason for performing a step is judged to be unnecessary, there is no need to do it. For example: a manual method of moving parts from process A to process B is replaced by a gravity-feed chute. The conveyance work is not improved — it is gone. Elimination removes the root of the waste rather than trimming its cost.
Combine. Work that cannot be eliminated is examined for separation or combination. Separation example: a visual inspection performed at the final process of a machining line takes about 15 seconds per cycle. That inspection is broken down and distributed across each forming process on the line — each operator inspects their own output. The centralized final inspection is eliminated as a result. Combination example: two team members work side by side; one cannot finish within task time while the other has waiting time every cycle. The group leader moves tasks from the overloaded worker to the worker with slack so both now equal task time.
Rearrange & Optimize. This direction makes work easier to perform. When a work combination based on Standardized Work results in processes being isolated from each other, large walking distances create waste; a partial layout change reduces the distance. Optimization also covers tooling design: in a paint shop, the tip profile of the sealer-gun nozzle was circular and could not reach certain areas satisfactorily. Floor employees used their resourcefulness to design a modified tip profile that solved the problem. Optimization means reducing unnecessary movement, reducing fatigue, and using the capability of people and equipment more fully.
Simplify & Standardize. In the Toyota Production System there is no kaizen without standardization. Random working speed, random procedures, random tools — random output and quality result. Costs go up; lead time and safety are affected. Standardization integrates the best method known at any point in time. It is collective resourcefulness. Once a kaizen improvement is confirmed, it must be written into the Standard Work and observed — otherwise the improvement evaporates and the next round of kaizen has no stable baseline to build from.
Synchronize. Synchronization maintains balance between the flow and timing of related processes. When operations are synchronized, the right parts are present and ready for processing at the right place and the right time. If one process finishes too early or too late, work according to task time is interrupted — irregular flow that creates overproduction, waiting, and quality problems downstream. Synchronization is the Just-in-Time concept applied at the method level: make what is needed, when it is needed, in the amount needed.
Build-in Quality. One of the two main pillars of TPS is in-station process control, which draws directly from the concept of jidoka. Jidoka means automatically stopping the line when any abnormality is detected in the part or the machine. Most machines are designed to run continuously regardless of whether the output is good. Under jidoka, the emphasis shifts to detecting abnormalities immediately and stopping at the source — preventing a bad part from reaching the next process. Primary emphasis always goes to eliminating the problem entirely; build-in quality is the backstop when a problem cannot be eliminated.
These six directions extend and sharpen the classic ECRS framework (Eliminate, Combine, Rearrange, Simplify). Synchronize and Build-in Quality reflect Toyota's particular emphasis on flow and on in-process detection — both of which are under-weighted in a generic ECRS checklist. Use all six when developing proposals.
5The implementation plan
Once the best proposal is chosen, it must be translated into an action plan before implementation begins. The action plan answers four questions for every item: what exactly needs to be done, who is responsible, by when, and what is the current status. Without these four answers in writing, "we agreed to do it" and "it was done" are rarely the same thing.
| Action Item | Responsible | Due Date | Status |
|---|---|---|---|
| What exactly needs to be done? | Who owns it? | By when? | Current condition |
| Relocate parts bin to operator's right side; update standard work | Team Leader | Nov 3 | In progress |
| Revise inspection instruction sheet to reflect in-station check | Process Engineer | Nov 7 | Scheduled |
| Confirm cycle time after method change; sign off standard work | Group Leader | Nov 10 | Pending Nov 3 completion |
Every kaizen plan is reduced to this table. Each row is one discrete action, owned by one person, with a specific due date and a running status update.
What to notice: "Responsible" names a person, not a department or a vague "team." A department cannot be held to a due date. The Status column is a live field — it changes at every review until the row closes.
The implementation plan is not a wish list. Each action item should be specific enough that a neutral observer could determine on the due date whether it was done or not. "Improve the process" is not an action item. "Relocate the parts bin to the operator's right side and update the standard work sheet" is.
6Worked example — grinding-machine defect
The following countermeasure table comes from an actual investigation into surface defects on a grinding machine. The problem was identified and the top causes selected in Step 2 (via Pareto analysis); Step 4 turns those suspected causes into a structured plan of investigation and action. Each row assigns one suspected cause to one person with a specific check date. The Finding column records what was actually discovered — it fills in as the work is done. This same example carries forward into Step 6, where the confirmed countermeasures are plotted on an effect-confirmation chart.
| # | Suspected Cause | Action Item | Responsible | Due | Finding |
|---|---|---|---|---|---|
| 1 | Dirt / contamination | Conduct daily 5S & PM | Tony (T/L) | 11/2 | Conducting daily — no issues found |
| 2 | Grinding wheel setup | Check setup procedure | Tony (T/L) | 11/4 | Checked OK |
| 3 | Manual offset function | Check offset settings | Tony (T/L) | 11/4 | Checked OK |
| 4 | Spindle bearing loose | Inspect spindle bearing | Ed (Maint) | 11/5 | Loose bearing cap found — tightened |
| 5 | Clamp / locator damage | Inspect clamp and locators | Ed (Maint) | 11/5 | Nothing abnormal |
| 6 | Grinding wheel balance | Check wheel balance | Tony (T/L) | 11/5 | Nothing abnormal |
| 7 | Incoming part dimensions | Measure incoming parts | Janet (QC) | 11/9 | Within spec |
| 8 | Poor material hardness | Measure material hardness | Janet (QC) | 11/9 | Within spec |
| 9 | Abnormal surface finish | Check surface-finish gauges | Janet (QC) | 11/9 | Within spec |
| 10 | Grinding conditions | Review grinding parameters | Mary (Eng) | 11/13 | Nothing abnormal |
| 11 | Coolant concentration | Measure coolant concentration | John (Maint) | 11/13 | Tanks contaminated — replaced |
| 12 | Wheel dressing | Check dressing procedure and condition | Mary (Eng) | 11/13 | Nothing abnormal |
Twelve suspected causes, each assigned to a responsible person with a specific check date. Two countermeasures yielded actual findings (rows 4 and 11); the rest were ruled out. The table documents both what was done and what was found.
What to notice: the investigation is cross-functional — team leader, maintenance, quality, and engineering each own their relevant rows. Responsibility follows domain knowledge, not hierarchy. Row 11 (contaminated coolant) and row 4 (loose bearing cap) are the confirmed root causes; both appear again in the Step 6 effect-confirmation chart.
Notice that most rows produced no finding. That is not wasted effort — ruling out causes is how you isolate the real ones. The discipline is to check systematically, record honestly, and not assume a cause is active just because it is plausible. Two causes were confirmed: a loose spindle-bearing cap and contaminated coolant. Those become the countermeasures to implement and confirm in Steps 5 and 6.
Section summary
Step 4 is the bridge between ideas and action. The first discipline is to generate multiple proposals — at least three — and evaluate each against quality, cost, delivery, and safety before committing to one. The joining-of-steel-sheets example (adhesive, caulk, bolts, weld, cord) illustrates the point: the space of options is almost always larger than the first idea suggests.
All proposals are either work kaizen (method and standardized-work changes; low cost; everyone participates) or equipment kaizen (machinery and automation; capital; specialist crew). Work kaizen takes precedence. Automating a flawed method produces a faster flawed method.
The six directions of kaizen — Eliminate, Combine, Rearrange & Optimize, Simplify & Standardize, Synchronize, Build-in Quality — are the moves available when developing proposals, listed in priority order. Elimination is always preferred; standardization is the condition that makes any earlier improvement permanent.
Once the best proposal is chosen, it is turned into an implementation plan with four columns: Action Item · Responsible · Due Date · Status. Each row names one person for one action with one date. The grinding-machine countermeasure table shows this in full: twelve suspected causes, each investigated by domain — two confirmed root causes found. Those confirmed causes move to Steps 5 and 6.