A good A3 fits on one page and changes something. Most hoshin A3s manage the first and miss the second: they find a root cause, list countermeasures and never say whether the plan itself should change. One extra section fixes that.
In this guide
Key Takeaways
- →Eight sections: the seven standard ones plus a plan-change section for hoshin use.
- →Section 8 is the shortest on the page and decides whether the A3 teaches the plan anything.
- →Each section on the left uses the one before it. A cause that doesn’t explain the data isn’t the cause.
- →A countermeasure with no cause it answers is the owner’s existing plan, relabelled.
- →Two weeks from flag to agreed A3 keeps findings inside one monthly cycle.
- →Overflowing one page is a sign the problem isn’t understood well enough to summarize.
The page
The A3 format comes from Toyota’s practice of working through a problem on a single sheet of A3 paper. John Shook’s Managing to Learn is a widely used guide to the thinking behind it. The layout below follows the standard sections, with one addition for hoshin use.
| Left side: understanding | Right side: acting |
|---|---|
| 1. Background: why this gap matters | 5. Countermeasures: what will be done |
| 2. Current condition: what is happening, with data | 6. Implementation plan: who, what, when |
| 3. Target: what good looks like, and by when | 7. Follow-up: how and when we’ll check |
| 4. Root cause analysis: why it is happening | 8. Plan change: what changes on the X-matrix |
Whether a gap deserves this page at all is decided by the triggers in When a Red Bowler Row Needs an A3 in Hoshin Kanri.
Section by section
1. Background
Which objective and row, why it matters, and which correlation on the matrix depends on it. Two or three sentences.
2. Current condition
Data, not opinion: the bowler trend, the leading indicator, where the gap is concentrated. A chart beats a paragraph.
3. Target
Where the row should be and by when. Usually the plan value, sometimes an interim recovery point.
4. Root cause
Use 5 whys, a fishbone or a Pareto chart to get past the first explanation. See Root-Cause Tools for Hoshin Kanri Gaps: 5 Whys, Fishbone, Pareto. Test the cause against the data in section 2.
5. Countermeasures
Each linked to a root cause, with its expected effect.
6. Implementation plan
Owners and dates.
7. Follow-up
What will be checked, when, and what result would show the countermeasure worked.
8. Plan change
The hoshin addition. Does this finding change a target, a correlation grade, a priority’s measure, or a priority itself? State which, or state “no change” explicitly.
Section 8 is the shortest section on the page and the one that decides whether the A3 teaches the plan anything.
A worked example: an A3, summarized
Here is a hypothetical A3, condensed to one line per section:
| Section | Entry |
|---|---|
| Background | Services share flat despite network priority at 85%; primary correlation at risk |
| Current condition | Attach rate up only in two of six territories; 70% of calls still dispatched from head office |
| Target | Attach rate in new territories at 30% by September |
| Root cause | Routing rules unchanged when centers opened; nobody owned the routing change |
| Countermeasures | Routing rules changed for all six territories; local call handling KPI per center |
| Implementation | Service operations lead; complete by 15 July |
| Follow-up | Check calls served locally and attach rate at August and September reviews |
| Plan change | Priority measure changed from centers opened to share of calls served locally; correlation kept as primary |
Common A3 mistakes
- Jumping to countermeasures. A left side written after the right side was decided.
- Current condition without data. If the gap can’t be shown, the cause can’t be tested.
- A blank section 8. The finding stays in the A3 and the plan repeats the mistake.
- More than one page. Overflow is a sign the problem hasn’t been understood well enough to summarize.
The left side in full
The summary table above condenses each section to a line. Here is how the left side of the same illustrative A3 reads when written properly.
Background
Services share of sales is the KPI for the breakthrough “services at 40% of sales by 2028”. This year’s target is 34%. The service network priority is graded primary against it, with an expected lag of two quarters after each center opens. Four of six centers opened by April; their lag windows have now closed.
Current condition
Services share is 31.4% against a plan of 32.6%. The leading measure, contract attach rate on new equipment in the new territories, is 23% against a plan of 30%. A Pareto of the shortfall by territory shows four of six territories account for about 90% of it. In those four, 70% of service calls are still dispatched from head office.
Target
Attach rate in the new territories at 30% by the end of September, measured monthly. Lagging check on services share after the December close.
Root cause
Ishikawa candidates: technician skills, contract terms, call routing, competitor pricing. Data test: the two territories with healthy attach rates had changed routing locally; the four lagging ones hadn’t. 5 Whys from “call routing” ends at: the network priority was defined as opening centers, not as changing how calls flow, so nobody owned the routing change.
Each section on the left uses the one before it. If the root cause doesn’t explain the current condition’s data, it isn’t the root cause yet.
What a coach asks
An A3 is reviewed before it commits anyone’s work. These are the questions an experienced coach works through, in order.
- Is the gap real? Does the current condition show it with data, not description?
- Is it located? Has a Pareto or similar narrowed where the gap actually is?
- Was the cause tested? Is there evidence for the chosen cause that rules out the alternatives?
- Does every countermeasure answer a cause? A countermeasure with no cause is the owner’s existing plan, relabelled.
- Will we know if it worked? Is there a leading check soon, and a lagging check after the expected lag?
- What changes on the matrix? Is section 8 filled in, even if the answer is no change?
Questions 3 and 4 catch most weak A3s. Who else needs to agree before the A3 commits work is covered in Who has to agree to an A3.
How long it should take
| Stage | Typical time |
|---|---|
| Current condition and data | Two to three days, mostly getting the data |
| Root-cause analysis | One to two working sessions |
| Countermeasures and plan | One session, with the people whose work is committed |
| Coach review and revision | A few days |
| Total, flag to agreed A3 | About two weeks |
Two weeks keeps the findings inside one monthly cycle. A3s that take two months usually lost time on the right side, negotiating countermeasures that the left side hadn’t justified.
The right side in full
Countermeasures
C1, answering the routing cause: change dispatch rules in all four lagging territories so calls go to the nearest center. Owner: service operations lead. C2, answering a secondary cause on contract terms: offer the multi-year contract in the new territories first. Owner: commercial director.
Implementation plan
C1: routing rules changed by 15 July; call-handling measure per center live by 31 July. C2: contract offer available in new territories by 1 September.
Follow-up
Leading check: share of calls served locally at 70% by the August review; attach rate at 30% by September. Lagging check: services share after the December close, compared with the two territories that had already changed routing.
Plan change
Change the network priority’s measure from centers opened to share of calls served locally. Keep the correlation to services share as primary; the two healthy territories support the mechanism. No target change proposed.
Notice what the right side doesn’t contain: any countermeasure without a cause it answers, or a check that would be judged before the lag has passed.
What a blank section 8 costs
Sections 1 to 7 cost about two weeks of an owner’s time. Section 8 costs about two sentences. Leave it blank and the two weeks buy a fix for one territory’s routing, while the matrix keeps measuring the priority by centers opened, so the same misdefinition produces the same gap next year and the reflection has nothing to learn from.
In the worked example, section 8 is the sentence that changes how the priority is measured. Without it, every other section is still correct and the plan is unchanged. That is the whole argument for the eighth box, and it is why it belongs on the page rather than in a follow-up email.
Among published alternatives for keeping that link, OnStrategy lists $200 per month for unlimited users. i-nexus publishes a free Starter tier for up to five users and $28 per user per month at its Standard tier. The relevant test is whether a plan change recorded in an A3 reaches the correlation it questions, or stops at the document.
Profit.co publishes two plans, Essentials and Enterprise, quoting against modules and scale rather than a public seat rate. Hoshin Kanri sits in Essentials alongside OKRs and Goals, Balanced Scorecard and Strategy Roadmap. Primary, secondary, partial: grading X-matrix correlations covers what a regrade has to mean, What drives the cost of hoshin kanri software lists the rest to total, and the 30-day free trial covers the two weeks this article budgets for one A3.
Writing an A3 in Profit.co
Profit.co builds the countermeasure A3 into the hoshin module, laid out as the sheet it will be read as: Plan sections on the left, Do, Check and Act on the right.
| Section of this article | How the product handles it |
|---|---|
| Background, current condition | Evidence blocks filled from the anchor KPI or initiative, the bowler row and gap, and the correlation, with an as-of date |
| Root cause | Ishikawa and 5 Whys editors; causes can go three levels deep; only the owner marks a root cause |
| Countermeasures | A table that references causes by ID; each countermeasure is an initiative, created or linked |
| Implementation plan | Tasks stored as milestones on the countermeasure initiative |
| Follow-up | Leading and lagging check measures; the lagging check opens only after the expected lag, and shows “not yet judged” until then |
| Plan change | The Act step: correlation decisions written to the edge’s strength history; target changes opened as catchball |
Two design choices follow the advice above. A missed check doesn’t get edited into success: it closes the revision with its verdict and opens the next one on the same gap. And review is consensus: everyone whose work the A3 commits agrees to the revision as sent, as described in Who has to agree to an A3.
An AI assistant in the Ishikawa and 5 Whys editors can suggest causes or challenge a chain of whys. Every suggestion is a proposal the owner applies or discards, applied content stays tagged as AI-suggested, and the AI can’t mark a root cause. The thinking stays the owner’s. More on the Hoshin Kanri software page.
Free Product Demo
Connect your A3s to the plan they question
A 30-minute walkthrough, tailored to your team.
Book a free demo
What happens in the call
How your teams set goals, track progress and review results.
Strategy, performance, projects and engagement on one platform, with AI.
A practical path to align teams, execute faster and drive results.
Questions this article answers
Background, current condition, target, root cause analysis, countermeasures, implementation plan and follow-up. Hoshin A3s should add a plan-change section.
A short statement of whether the finding changes a target, correlation grade, priority measure or priority on the X-matrix, or explicitly that nothing changes.
One page. If it overflows, the problem usually isn’t understood well enough to summarize.
Whether the cause was tested with evidence that rules out alternatives, and whether every countermeasure answers a named cause. A countermeasure with no cause is the owner’s existing plan, relabelled.
About two weeks from flag to agreed A3, which keeps the findings inside one monthly cycle. A3s that take two months usually lost time negotiating countermeasures the left side had not justified.
The finding stays inside the A3 and the plan repeats the mistake. Even “no change” should be stated explicitly, so the decision is recorded rather than skipped.
