Implementation Guide
A3 Problem Solving.
An A3 is a one-page problem-solving format that walks one owner through PDCA: the left side builds understanding of the problem, the right side commits to countermeasures and follow-up. Its real power is not the paper but the coaching conversation at every section. An A3 written alone the night before a review is just a small report, and most A3s fail because the left side was rushed.
What an A3 is, and what it is for
An A3 is a problem-solving story told on one sheet of paper, named for the paper size: 297 by 420 millimeters, about 11 by 17 inches. Toyota standardized on the format not because the size is magic but because the constraint is. One page has no room for everything that happened, so the owner has to decide what matters, and a reader can see the whole argument, from problem to result, without turning a page.
The layout is PDCA made visible. The left side is Plan: background, current condition, target condition and root cause analysis. The right side is Do, Check and Act: countermeasures, implementation plan, follow-up and results. Reading an A3 left to right is reading someone's thinking in the order the thinking should have happened. The left side is understanding; the right side is action; and the quality of the right side is decided entirely by how honestly the left side was built.
Which is why the A3 is a coaching artifact, not a report format. Its power is the conversation between the owner and a coach at each section: the questions asked, the verification demanded, the trips back to the process in between. An A3 written alone the night before a review carries the same boxes and none of the value. It is a small report. The method develops two things at once, a solved problem and a better problem solver, and the second one is the reason Toyota kept the format for decades.
Where A3 sits in the transformation roadmap
On the TeamGuru deployment roadmap, A3 belongs to the structured problem solving practice in the Run Daily stage. It assumes a daily management system already exists and is surfacing problems faster than huddles can absorb them. On the problem-solving ladder, A3 is the middle rung: bigger than a 5 Whys huddle at the board, lighter than a customer-facing 8D. It exists for the problems that survive the daily meeting: the ones that cross shifts, recur for months, and never quite become anyone's job to kill.
- Before Daily Management
- You are here Structured Problem Solving
- After Kaizen
- After Obeya & Management Reviews
When to use which tool
Rigor should match consequence. Running an A3 on every deviation produces format fatigue and shallow A3s; leaving chronic problems at the huddle level produces the same 5 Whys every Tuesday. A working selection rule:
| Tool | Reach for it when | Typical effort | Example |
|---|---|---|---|
| 5 Whys at the board | A same-day issue with one causal chain, and the cause is likely visible at the process. | 30 to 60 minutes, same day | A missed shift plan caused by a material shortage |
| A3 | A chronic problem crossing shifts or functions, cause unknown, worth weeks of one owner's attention. | 2 to 6 weeks, in short coaching sessions | Changeovers overrunning the standard for months |
| 8D | A customer is affected, or a formal escape demands containment, evidence and a report. | Containment in days, corrective action typically within 30 days | Defective parts shipped to a customer |
| Improvement project | The solution is already known and verified; only execution remains. | Whatever the plan requires | Installing the staged tooling an A3 already proved |
If your organization runs a formal corrective-action discipline with effectiveness checks, see the RCCA guide: an A3 is a natural vehicle for the analysis inside such a case, and the two are complementary, not competing.
How to write one: the left side first, and slowly
The order of writing is the method. Countermeasures are the most enjoyable section to write and the most dangerous place to start, because a countermeasure chosen before the current condition is understood will bend the whole left side into its justification. Write top left to bottom right, and let each section earn the next. These rules keep the left side honest:
- Go see before writing anything. The current condition is drafted standing at the process, not recalled at a desk. If the A3 is about changeovers, watch changeovers.
- Data over adjectives. Often, unstable and too long are placeholders for measurements not yet taken. Write 84 changeovers averaged 68 minutes against a 47-minute standard, not changeovers keep running late.
- Draw the current condition. A simple sketch of the flow with the problem located in it beats three paragraphs of description. If you cannot draw it, you do not understand it yet.
- Make the target a sentence with a number and a date. Metric, value, deadline. Everything on the right side will be judged against it.
- Verify every cause at the process before it earns a place on the page. A cause agreed on in a meeting room is a hypothesis, not a cause.
- Write countermeasures as experiments. Each maps to one verified cause and predicts its effect. A prediction can turn out wrong, which is exactly what makes the follow-up section meaningful.
The right side is quicker to write and easier to get wrong. Owners are names, never teams or departments. The first implementation step is small enough to start this week, because a plan that starts next quarter is a plan that starts never. And the follow-up dates go into real calendars before the plan begins, agreed with the coach, so checking results is an appointment rather than an intention.
The coaching rhythm
An A3 has two people attached: an owner close to the problem, and a coach, usually the owner's manager or a trained mentor. They meet for about 20 minutes per section, at the board or at the process, spread over days or weeks. Short and frequent beats long and rare: between sessions the owner goes back to the process to measure, observe and verify, and that is where the analysis actually improves. Eight sections reviewed in one hour-long meeting is a presentation, not coaching.
The coach asks and never writes. This deserves defending, because it is the most violated rule in the method. The moment the coach picks up the pen, ownership transfers: the owner stops investigating the problem and starts guessing what the coach wants to read, and the A3 stops developing anyone's judgment. A coach who supplies answers gets compliance and a dependent team; a coach who supplies questions gets thinking. The questions are simple and repeatable: how do you know, what did you see, what effect do you expect, how will you check. The payoff compounds: the owner's second A3 needs fewer questions than the first.
The pace of an A3 is set by verification, not by writing. Two to six weeks is a normal life for a chronic problem, because confirming a cause at the process, trialing a countermeasure and reading the results takes calendar time. An A3 completed in an afternoon almost always carries an unverified left side, and an unverified left side produces a right side of guesses.
The eight sections, one by one
Every section answers one question, and every section has a well-worn way to fail. The coaching question column doubles as a self-check when no coach is available yet: ask it of your own draft before anyone else does.
| Section | The question it answers | What good looks like | The coach asks | Classic mistake |
|---|---|---|---|---|
| 1. Background | Why does this problem matter, and why now? | Two or three sentences linking the problem to a business or customer effect, with the scale quantified. | Why should anyone outside your area care about this? | A history lecture, or the title restated in more words. |
| 2. Current condition | What is actually happening today? | Data measured at the process, a simple drawing of the flow with the problem located in it, and the gap stated as a number. | How do you know? Show me where this number comes from. | Adjectives instead of measurements: often, unstable, too long. |
| 3. Target condition | What outcome, how much, by when? | One metric, a value and a date, tied directly to the gap in the current condition. | How will we know you have succeeded? | A direction without a number, or a number without a date. |
| 4. Root cause analysis | Why does the gap exist? | Causal chains verified with evidence at the process, using 5 Whys or a fishbone, ending at a system cause. | How did you verify this cause, and what ruled out the others? | A favorite solution dressed up as a cause. |
| 5. Countermeasures | What will you try against each verified cause? | Experiments mapped one-to-one to causes, each with an expected, quantified effect. | What effect do you predict, and by when? | A shopping list of good ideas unconnected to the analysis. |
| 6. Implementation plan | Who does what, by when? | Named owners, dates, and a small first step that can start this week. | What happens first, and who exactly does it? | Owners that are departments and dates that are quarters. |
| 7. Follow-up | When and how will results be checked? | Check dates in real calendars, the metric named, the review forum agreed. | When do we look at the data together? | Written once at the start and never opened again. |
| 8. Results | What actually happened against the target? | The same metric as the target, the honest gap, what was learned, and what the standard says now. | What did you learn, and what changed in the standard? | Success declared in prose while the section stays empty of numbers. |
A filled example: the late changeover A3
The numbers below are illustrative but internally consistent, for the same 450-person components manufacturer used across these guides: 456 units per day, takt of 118 seconds, a 47-minute machining changeover standard. The A3 is condensed to one line per section; a real one carries the timeline sketch of an observed changeover, the causal chains and the data behind each line. The discipline is the same.
A3: Machining changeovers overrun the 47-minute standard
Owner: Machining supervisor
Coach: Value stream manager
Status: Closed, week 8
Left side: understanding (Plan)
1. Background
Machining feeds assembly at a takt of 118 seconds and runs 4 to 6 changeovers a day. Overruns cost one to two hours of capacity daily and force overtime to hit 456 units.
2. Current condition
84 changeovers measured last month: average 68 minutes against the 47-minute standard. Timing 12 of them showed the machine idle for the first 15 to 20 minutes while tools, fixtures and programs were gathered.
3. Target condition
Average changeover at or below the 47-minute standard, with 9 of 10 changeovers within standard, in 8 weeks.
4. Root cause analysis
Preparation is not part of the changeover standard: gathering tools, fixtures and programs starts only after the machine stops, and preparation sits in nobody's standard work. Verified by observing five changeovers end to end.
Right side: action (Do, Check, Act)
5. Countermeasures
Stage a changeover cart with tools, fixture and first-piece gauges during the last hour of the previous run, guided by an external-preparation checklist. Expected effect: 15 to 20 minutes off every changeover.
6. Implementation plan
Setter and supervisor build the cart and checklist in week 2, trial on 10 changeovers in weeks 3 and 4, train both shifts by week 5. Preparation added to the setter's standard work.
7. Follow-up
Changeover minutes reviewed weekly at the tier 2 board for 8 weeks; the supervisor audits preparation twice a week.
8. Results
Week 6: average 49 minutes, 87 percent within standard. Two outliers traced to a fixture shared with another line, now scheduled. Standard updated to include preparation.
Illustrative example. A real A3 attaches the observation data and sketches behind each line.
Two things in this example are worth copying. The countermeasure targets the system cause, preparation being in nobody's standard work, rather than the setter's diligence. And the case ends in a changed standard plus a handoff: once changeovers reliably hit 47 minutes again, cutting the standard itself from 47 to 18 minutes became a SMED project with a stable baseline to improve from. An A3 that restores the standard and a project that improves it are different jobs, and confusing them sinks both.
Common mistakes
A3 programs rarely fail loudly. They fail quietly, into template compliance: the boxes get filled, the reviews get held, and nothing about how the plant thinks changes.
What bad looks like
- The template is filled in, in order, in one sitting, the night before the review
- The CI office ghostwrites A3s that line managers then present
- The right side is written first and the left side reverse-engineered to justify it
- The follow-up and results sections stay empty forever, and nobody notices
- Ten open A3s per owner, none of them moving
What good looks like
- One owner, one coach, one page, one problem worth weeks of attention
- A current condition built from data measured at the process, with a drawing
- Every countermeasure maps to a verified cause and predicts its effect
- Follow-up dates sit in calendars and produce an honest results entry
- The closed A3 ends in a changed standard, not just a closed case
After the A3: making countermeasures stick
A verified countermeasure is a discovery about how the process should run, and it survives only if the standard changes with it. The last act of a real A3 is updating standardized work and training the change, so the next audit checks the new method instead of the old one. Countermeasures that proved themselves in one area are also raw material for the improvement system: a fix worth keeping is usually a fix worth spreading.
A3 thinking also scales upward. The same left-side-then-right-side discipline that handles a changeover problem carries strategy problems at plant level, which is why hoshin planning is often described as A3 thinking applied to the year. A leadership team that coaches A3s well tends to deploy strategy well, because both habits are the same habit: understanding before action.
The practical failure point between coaching sessions is memory: who owed what, which data was supposed to be collected, when the follow-up was due. This is where teams running the method digitally use TeamGuru: the root cause analysis use case holds the chains, evidence and countermeasures in one record, and actions carry the right side's owners and dates, so follow-up happens on a schedule instead of by recollection. The page stays one page; the system around it stops leaking.
