Skip to content

Method guide · Root cause analysis and problem solving

A3 problem solving: the A3 report format, the steps and a filled example

Short answer

A3 problem solving is a lean method that takes one problem from background to confirmed result on a single sheet of A3 paper. The left half of the sheet describes the problem and its root cause; the right half holds the countermeasures, the plan and the check of results, in plan, do, check, act order.

An A3 is a one-page report with seven boxes: background, current condition, goal, root cause analysis, countermeasures, implementation plan and follow-up. One person owns it, writes it in pencil, takes it to the floor and to everyone the fix affects, and rewrites it after each conversation. A coach reviews it box by box. The page limit forces facts over opinions, and the finished sheet tells anyone who picks it up what went wrong, why, what was done and whether it worked. A3 thinking is the habit behind the sheet: understand the problem before choosing the fix, and prove the fix with numbers.

Last reviewed:

What is A3 problem solving?

A3 problem solving is a structured way to work a problem from first report to a result you can prove, written up on one A3 sheet. The sheet follows plan, do, check, act: most of it is plan, because that is where teams usually rush. It is used for problems too big for a quick fix at the line and too small, or too local, for a full Six Sigma project.

The name is the paper. A3 is the international size 297 x 420 mm, close to the 11 x 17 inch tabloid sheet used in North America. Toyota made the one-page report its standard way to write up problems, proposals and progress, and the size stuck as the name of the method. Two books describe it from inside that practice: Understanding A3 Thinking by Durward K. Sobek II and Art Smalley (Productivity Press, 2008), and Managing to Learn by John Shook (Lean Enterprise Institute, 2008). Our pages on Understanding A3 Thinking and Managing to Learn summarize each one.

The word A3 is used three ways. The A3 report is the sheet. A3 problem solving, sometimes called the A3 process or A3 methodology, is the sequence of steps the sheet records. A3 thinking is the discipline under both: go and see, separate facts from guesses, test causes before choosing fixes, and agree the plan with the people it affects. A plant can fill in A3 forms without any of the thinking, which is the most common way the method fails.

The A3 format: the seven boxes

Layouts vary by company, but nearly every A3 report has the same seven boxes. The left side is the story of the problem. The right side is what the team will do about it and how they will know it worked. A title, the owner, the coach and the date sit across the top.

An A3 from first box to close

  1. 1Plan

    Background

  2. 2Plan

    Current condition

  3. 3Plan

    Goal

  4. 4Plan

    Root cause analysis

  5. 5Plan

    Countermeasures

  6. 6Do

    Implementation plan

  7. 7Check

    Follow-up

  8. 8DecisionAct

    Did the result meet the goal?

    Yes:
    standardize, share and close the A3
    No:
    back to root cause analysis
Boxes 1 to 4 are the left side of the sheet, 5 to 7 the right side. A result that misses the goal goes back to the cause analysis.
The seven boxes of an A3 report
BoxSideWhat goes in itThe coach's question
1. BackgroundLeft (plan)Why this problem matters now, in safety, quality, delivery or cost terms, and who feels itWhy should anyone spend time on this?
2. Current conditionLeft (plan)What happens today, from the floor: a chart, a sketch of the process, the gap between the standard and realityDid you see it yourself, and where are the numbers?
3. GoalLeft (plan)A measurable target and a date, ideally tied to the gap in box 2How will we know when this A3 is finished?
4. Root cause analysisLeft (plan)Fishbone, 5 Whys, is / is not or a Pareto: how the team got from symptom to cause, and how each cause was checkedHow do you know that is the cause and not a guess?
5. CountermeasuresRight (plan)One or more actions per confirmed cause, with the effect each is expected to haveWhich cause does each action remove?
6. Implementation planRight (do)Who does what by when, and the status of each actionWho has agreed to this, and what is late?
7. Follow-upRight (check, act)The result against the goal and the prediction, what becomes standard, and where else it appliesDid it work, and how do we keep it?

Some A3s split box 7 into confirmation of results and follow-up actions, or add a separate box for the target condition; the order does not change. The free [A3 problem solving template](/templates/a3-problem-solving-template) has these seven boxes as a PDF and an Excel file with a worked example.

Write the left side before the right side, and do not put a solution in the title. "Porosity rework on weld cell 3" names a problem; "New gas hoses for cell 3" names an answer before anyone has looked. Keep the boxes small on purpose: if the current condition does not fit, the team has not yet decided which facts matter.

How to do A3 problem solving, step by step

  1. 1

    Pick a problem that deserves an A3, and one owner

    A gap against a standard that keeps coming back, costs real time or money, or needs more than one department to fix. One person owns the sheet. On a shop floor that is often the engineer or supervisor closest to the problem, with a manager as coach.

  2. 2

    Go to the floor and draw the current condition

    Watch the process, collect the numbers, and sketch what happens. Split the data by shift, machine, part or operator until the problem has a shape. A Pareto chart of reasons is usually the first picture on the sheet.

  3. 3

    Set the goal

    A number and a date, such as rework on cell 3 at or below 2.0% for four weeks in a row by week 12. Where one machine or shift already runs well, its level is a fair first goal for the others.

  4. 4

    Find the root cause and test it

    List the candidate causes, then check each one on the floor or in the data and strike out the ones the facts rule out. Dig into the survivor with 5 Whys. A cause is confirmed when removing it stops the problem.

  5. 5

    Propose countermeasures and get agreement

    Write at least one countermeasure per confirmed cause and predict what it will do to the number. Take the draft to everyone the change affects, listen, and rewrite. Lean teams call this round of conversations nemawashi.

  6. 6

    Run the plan

    Each action has an owner and a date. Review the plan in the team's regular huddle so late actions are seen the day they slip, not at the next monthly review.

  7. 7

    Check the result against the goal and the prediction

    Measure the same thing the same way as the baseline. If the result misses, go back to the cause analysis: either a cause was wrong or there was a second one.

  8. 8

    Standardize, share and close

    Change the work instruction, the check sheet or the maintenance plan so the fix outlives the project. Copy it to similar lines (yokoten). Then close the A3 and pick the next problem from the new Pareto.

A first A3 usually takes a few weeks of elapsed time and several drafts. That is normal. The drafts are where the learning happens, and the A3 thinking post explains why a sheet finished in one sitting is usually a weak one.

A3 example: porosity rework on a robot weld cell

This A3 is illustrative: a realistic problem with made-up numbers, filled in box by box as it would be on the sheet. Cell 3 in a fabrication plant has two welding robots, A and B, that weld the same steel base frame, 1,600 frames a week. The owner is the cell's process engineer; the coach is the production manager.

Box 1, background

Rework on cell 3 is the largest internal quality loss on the frame line. Each reworked frame is ground out and welded again by hand, about 20 minutes of a skilled welder's time, and late frames make the paint line wait.

Box 2, current condition

Over four weeks the cell welded 6,400 frames and sent 240 to rework, a rate of 3.75%. A Pareto of the rework tags shows porosity, gas pockets in the weld, as two thirds of it. Split by robot, almost all the porosity is on robot B, although both robots weld the same frame with the same wire from the same parts rack.

Illustrative rework on weld cell 3, four weeks
Rework reasonFramesShare of rework
Porosity16066.7%
Spatter4016.7%
Undercut2410.0%
Other166.7%
Total240100% (3.75% of 6,400 frames)

Check: 160 + 40 + 24 + 16 = 240, and 240 / 6,400 = 0.0375, so 3.75%. The other way: 6,400 x 3.75% = 240. Porosity is 160 / 240 = 66.7% of the rework and 160 / 6,400 = 2.5% of all frames.

Illustrative rework on weld cell 3, four weeks, as a Pareto chart

Example numbers
  • Porosity: 160, 67% of the total, cumulative 67%.
  • Spatter: 40, 17% of the total, cumulative 83%.
  • Undercut: 24, 10% of the total, cumulative 93%.
  • Other: 16, 7% of the total, cumulative 100%.

Bars for porosity 160, spatter 40, undercut 24 and other 16 frames, with porosity alone at two thirds of the rework.

Porosity is 160 of 240 reworked frames, two thirds, so the A3 narrows to porosity before it looks at causes.
Illustrative porosity by robot, same four weeks
RobotFrames weldedPorosity reworkPorosity rate
Robot A3,200240.75%
Robot B3,2001364.25%
Cell 36,4001602.5%

Check: 24 / 3,200 = 0.0075 and 136 / 3,200 = 0.0425; 24 + 136 = 160. Robot B makes 136 / 160 = 85% of the porosity on half of the frames.

Box 3, goal

Rework on cell 3 at or below 2.0% of frames for four weeks in a row by the end of week 12. The number comes from robot A: if robot B ran at robot A's 0.75% porosity, the cell would have 6,400 x 0.75% = 48 porosity frames instead of 160, and rework of 48 + 40 + 24 + 16 = 128 frames. Then 128 / 6,400 = 2.0%.

Box 4, root cause analysis

The team listed candidate causes on a fishbone and checked each one against the fact that only robot B is affected.

Illustrative cause checks for porosity on robot B
Candidate causeHow it was checkedResult
Wire lotBoth robots feed from the same wire lotRuled out: same wire, different result
Oil or rust on the steelBoth robots take frames from the same press and rackRuled out: same steel, different result
Draft from the dock door near robot BOne shift with the door kept shut: 6 porosity frames out of 140 on robot B, about 4.3%Ruled out: no change from 4.25%
Shielding gas flow at the torchFlowmeter held on the nozzle of each torch; the weld procedure calls for 15 to 20 litres a minuteConfirmed: robot A 17, robot B 8, while both regulator gauges read 18

Gas leaves the regulator at the right rate on both robots, so robot B is losing it on the way to the torch. A soapy-water test on robot B's gas hose showed a leak where the hose rubs against the robot arm.

A 5 Whys from the confirmed cause: porosity on robot B, because too little shielding gas reaches the weld. Too little gas, because the hose leaks where it rubs on the arm. It rubs, because the clamp that holds the cable bundle clear of the arm was not refitted when the torch was replaced eleven weeks ago. Not refitted, because the torch change procedure has no step for it. And nobody saw the low flow, because the daily check reads the regulator gauge, which still showed 18. Two causes to remove: a missing step in the torch change procedure, and a daily check that measures in the wrong place.

The team turned the cause off to prove it: with a new hose and the clamp refitted, robot B read 17 litres a minute at the nozzle and had 3 porosity frames in its next 400, which is 0.75%, the same as robot A.

Boxes 5 and 6, countermeasures and plan

Illustrative countermeasures and plan for weld cell 3
CountermeasureCause it removesOwnerDue
Replace robot B's gas hose and refit the cable clampThe leaking hose (a correction, done first)Maintenance technicianDay 1, done
Add "refit the cable clamp, then read flow at the nozzle" to the torch change procedureThe missing stepMaintenance plannerWeek 6
Change the daily check to read gas flow at the nozzle, with the 15 to 20 range printed on the check sheetThe check in the wrong placeCell 3 team leaderWeek 6
Check clamps and nozzle flow on the other five robot cellsThe same risk elsewhere (yokoten)Maintenance plannerWeek 7

Prediction written on the sheet before the change: porosity on robot B falls to robot A's level and cell rework falls to about 2.0%.

Box 7, follow-up

Illustrative follow-up against the goal
Four-week periodFrames weldedReworkedOf which porosityRework rate
Weeks 1 to 4 (baseline)6,4002401603.75%
Weeks 7 to 10 (after)6,40011844 (22 on each robot)1.84%

Check: 118 / 6,400 = 0.0184, so 1.84%, under the 2.0% goal. The 118 is 44 porosity + 38 spatter + 22 undercut + 14 other. Porosity per robot is 22 / 3,200 = 0.69%, at or below robot A's old 0.75%. The small falls in spatter and undercut are within normal week-to-week noise, so the A3 does not claim them.

What the result is worth: 240 minus 118 is 122 fewer reworked frames in four weeks, at about 20 minutes each, so 2,440 minutes, roughly 40.7 welder hours. Checked the other way: 40.7 hours x 60 = 2,440 minutes, and 2,440 / 20 = 122 frames.

The goal is met, but the A3 stays open for a second four weeks before it closes, because one good month can be luck. The yokoten check found a loose clamp on one more robot. Spatter is now the top reason on the cell's Pareto, which makes it the next problem to pick up, as a quick kaizen or a new A3. For three more filled A3s, one of each kind with a coach's note on every box, see A3 examples.

Types of A3: problem solving, proposal, status and strategy

Sobek and Smalley describe three kinds of A3 report, each answering a different question. Many companies add a fourth for strategy.

The kinds of A3 and when each is used
Kind of A3The question it answersTypical shop-floor use
Problem-solving A3Why does this keep happening, and what fixes it?Scrap, rework, a recurring breakdown, a safety near miss that repeats
Proposal A3Should we do this, and which option is best?A new fixture, a layout change, a capital request, a changeover project
Status A3Is the project on track against its goal?A monthly review of a long improvement project
Strategy A3What must this site or department achieve this year, and how?The yearly plan that feeds hoshin kanri and the plant's improvement projects

The strategy A3 is a later use of the same page discipline rather than one of the book's three; see [hoshin kanri](/hoshin-kanri) for how yearly goals cascade to projects. The [A3 report post](/blogs/a3-report-project-management) covers using an A3 to run a project.

A3 vs 8D vs PDCA vs DMAIC

A3 is PDCA laid out on a page. 8D and DMAIC are longer, more formal sequences for particular kinds of problem. The right choice depends on who will read the result and how big the problem is.

A3 compared with 8D, PDCA and DMAIC
A38DPDCADMAIC
What it isOne-page report and process that follows PDCAEight disciplines (D1 to D8, with D0 to prepare) in a set report formatThe four-step improvement cycleFive-phase Six Sigma project method
Best forAn internal problem that takes weeks, owned by one personA customer complaint or quality escapeTesting a change quickly, often several timesA chronic, data-heavy problem over months
ContainmentOptional, inside the countermeasuresRequired early, in D3Not a set stepNot a set step
Who reads itYour own managers and teamsThe customer and the quality teamThe team running the testProject sponsor and finance
OutputOne sheet: problem, cause, fix, resultAn 8D report the customer acceptsA decision: adopt, adjust or drop the changeA project storyboard and verified savings

The steps of an 8D overlap with an A3: D2 is the current condition, D4 the root cause, D5 and D6 the countermeasures, D7 the standardization. The [8D post](/blogs/8d-problem-solving-eight-disciplines-manufacturing) explains each discipline and the [8D report template](/templates/8d-report-template) is free.

Use an A3 when your own plant will read the result. Switch to an 8D when a customer asks for one. A problem that has beaten two A3s, with months of data and many causes, may need a DMAIC project; Lean Six Sigma teams often summarize that project on one page too, on a DMAIC storyboard. The problem solving methods guide has a decision table for the full set, and the PDSA cycle guide explains the cycle the A3 follows.

A3 thinking and the coach

In Managing to Learn, Shook tells the story of a young manager and his boss working through one A3 over many drafts. The point of the book is that the sheet is a tool for coaching: the boss does not fix the problem, he asks questions until the owner can see the gaps in his own thinking. That is the part of A3 that most plants skip.

A coach reviewing an A3 asks the same few questions every time. Print them on the back of the template:

  • Did you see the problem yourself, on the floor, and where are the numbers?
  • Is the problem statement free of a solution?
  • What else could cause this, and what did you check to rule each one out?
  • If we remove this cause, does the problem stop? Have you tried it?
  • Which cause does each countermeasure remove, and what do you predict will happen to the number?
  • Who will this change affect, and have you talked to each of them?
  • How will we know in three months that the fix is still in place?

Thin answers show up as thin boxes, which is why an A3 is a good way to coach a new engineer or supervisor. A gemba walk is the natural place to review one: what is a gemba walk covers how leaders go and see.

Common A3 mistakes

  • Starting with the answer. The title or the background already names the fix, and every box after it is written to justify it.
  • A current condition with no numbers. Without a baseline there is no goal, no prediction and no way to say the A3 worked.
  • Causes agreed in a meeting room. Each cause on the sheet should say how it was checked: in the data, by a trial, or by turning it off and on.
  • Countermeasures that do not match a cause. If an action does not remove a confirmed cause, it is a guess, however sensible it sounds.
  • Writing the A3 alone at a desk and presenting it finished. The conversations with operators, maintenance and the next shift are the method, not a step after it.
  • Closing at implementation. The follow-up box is where you find out whether the problem is solved; many A3s stop the week the actions are done.
  • Treating the template as the method. A neat form with weak thinking is worse than a rough sheet with the right facts.

Most of these show up as soon as a coach asks the questions above. The A3 problem solving post goes through a ten-step version of the process with more on each step.

A3 template and tools

Start on paper or a whiteboard in pencil, because the first drafts change a lot. The free A3 problem solving template is a PDF that prints at A3 or 11 x 17 with a worked example on page 2, and an Excel version that works out the defect rate, flags late actions and shows whether each follow-up period met the goal.

For box 2 and box 4, the free Pareto chart maker builds the Pareto from a list of reasons, and the fishbone diagram maker lays out the candidate causes. The root cause analysis guide explains which cause tool fits which problem, and the fishbone diagram and Pareto chart guides go deeper on each.

In LeanSuite, Lean Studio has an A3 board with the seven boxes on one landscape canvas, and a tag raised on the floor that keeps coming back can be escalated into a Kaizen or A3 project in one click, keeping the tag's history. Kaizen & Projects then tracks the project through to results. If you are comparing digital tools, the best A3 problem solving software lists the options.

Continuous Improvement Software

Run A3 problem solving in LeanSuite

Use A3, DMAIC, PDCA, PDSA or just-do-it templates, or define your own method and phases.

FAQ

Questions about a3 problem solving

  1. A3 problem solving is a lean method that works one problem from background to a confirmed result on a single A3 sheet. The left side holds the background, current condition, goal and root cause; the right side holds the countermeasures, plan and follow-up. It follows plan, do, check, act, and one owner writes it with a coach reviewing it.

  2. A3 is the paper size the report is written on: 297 x 420 mm, close to an 11 x 17 inch sheet. Toyota used one sheet of that size for problem reports, proposals and status reports, and the size became the name of the method.

  3. The seven boxes, in order: background, current condition, goal, root cause analysis, countermeasures, implementation plan and follow-up. The first four describe the problem and its cause. The last three say what will be done, who does it by when, and whether the result met the goal.

  4. It is the practice Toyota developed of writing a problem, a proposal or a project status on one A3 sheet, in PDCA order, and using the sheet to coach the person who owns it. Durward Sobek and Art Smalley describe it in Understanding A3 Thinking and John Shook in Managing to Learn, both published in 2008.

  5. A3 thinking is the discipline behind the sheet: go and see the problem, separate facts from guesses, confirm the cause before choosing the fix, predict the result, and agree the plan with the people it affects. A plant can fill in A3 forms without it, and that is the most common way the method fails.

  6. An A3 is a one-page report for internal problems, owned by one person and read by your own managers. An 8D is a longer report in eight disciplines, used when a customer asks for one after a quality escape, with containment required early in D3. The thinking overlaps: both define the problem, confirm the root cause and verify the fix.

  7. An A3 is PDCA written on one page. Background, current condition, goal, root cause and countermeasures are the plan; the implementation plan is the do; the follow-up box is the check and act. PDCA is the cycle, and the A3 is a format and a coaching routine for running it.

  8. A3 comes from lean, not Six Sigma, but Lean Six Sigma programs use it in two ways: as the method for problems too small for a full DMAIC project, and as a one-page summary of a DMAIC project for reviews. The thinking is the same: define the gap, find the cause, fix it and prove it.

  9. It is a blank A3 sheet with the seven boxes laid out, the problem side on the left and the action side on the right, so the owner fills in the story in order. A good template also has room for a chart in the current condition and a table that checks the follow-up against the goal.

  10. No. It started in manufacturing at Toyota, but hospitals, offices and software teams use the same one-page format for problems and proposals. The boxes stay the same; only the data in the current condition changes.

Or browse the topics, the full library and the lean glossary.

Rather see it on your own problem?

Bring the problem you are stuck on, and we will show you on a call how LeanSuite handles it.

Pass it on

We want LeanSuite to be the best place on the internet for lean help. If this was useful, send it to someone on your team or in your network who needs it.