The situation

Cell 4 is the bottleneck again. Heat treat waits for shafts, overtime grows, and the weekly meeting has a nine-slide deck about it. Nobody remembers slide three. The plant manager asks one question: “What is the problem, and what are we doing about it?”

An A3 answers that question on one sheet of paper.

What an A3 is for

An A3 report is a one-page story about a problem. The name comes from the paper size: A3 is 297 × 420 mm, about two A4 sheets side by side. The page has boxes in reading order: background, current condition, target, causes, countermeasures, plan, follow-up.

An A3 does two jobs:

  • It makes the author think clearly, because the page is small.
  • It gives a manager or a neighbour a way to review the work in five minutes and ask good questions.

Terms used here:

  • Countermeasure: an action aimed at a verified cause.
  • Takt time: the pace at which a part must be finished to meet demand.
  • Gemba: the real place where the work happens.

A short history

The A3 comes from Toyota and the Toyota Production System. The usual story is that Taiichi Ohno wanted problems, proposals, and status reports on a single sheet, instead of long memos. We could not confirm that story from a primary source, so treat it as common lore. What is documented: John Shook explained the A3 management process to a Western audience in his 2008 book Managing to Learn, published by the Lean Enterprise Institute.

Where it applies, and where it does not

Use an A3 for:

  • scrap, yield, or bottleneck problems that cross shifts or departments;
  • improvement projects (kaizen);
  • briefing a manager who needs the problem, cause, and request on one page;
  • agreeing on a change with production, tooling, quality, and maintenance before it starts.

Do not use an A3 for:

  • a customer complaint where the customer demands a formal 8D format (use the 8D guide);
  • a small hiccup that the morning stand-up can solve;
  • a multi-year investment that needs a full project charter.

How to do it

Our example is a fictional plant, Silesia Gear & Axle. The part is pinion shaft SHAFT-104. Pawel facilitates. Ewa leads the shift. Jan is the maintenance technician.

  1. Title and background. Say why the problem matters. Cell 4 holds back heat treat, and overtime costs about 2,400 EUR a week.
  2. Current condition. Measure at the gemba. Show numbers and a sketch. Average cycle time is 72 s against a takt of 55 s. Operators spend 22 s filing burrs on the spline edges.
  3. Target. Make it measurable, with a date. Cycle time of 53 s or less by 30 July, without extra labour.
  4. Root cause. Ask why until you reach a process cause. Use a 5-Why or a fishbone. Mark each cause as observed or assumed. Inserts wear early because coolant pressure is 1.2 bar instead of 3.5 bar. An inline filter is clogged, and nobody cleans it on a schedule.
  5. Countermeasures. One action per verified cause. Fit a quick-change duplex filter. Add an in-cycle brush deburring tool at turret station 8.
  6. Plan. Who does what, by when. Jan fits the filter in week 24. Ewa updates the CNC program in week 25. A trial of 500 pieces runs in week 26.
  7. Follow-up. Set dated checks. Plot weekly cycle time on the SQCDM board for six weeks. Compare with the target in box 3. Record the result, and any surprises.

Read the page as a story. A reader should move from “why this matters” to “what we will check” without asking you anything.

Three ways it goes wrong

1. The decision came first. The team already wants a new machine, so the A3 is written backwards to justify it. The form becomes a sales brochure. If the answer is fixed, say so, and write a proposal instead.

2. Opinion over evidence. The current condition and causes are written at a desk. No one measured a part or watched a cycle. Go and see. Replace “I think” with “I measured”.

3. Box 7 left empty. The countermeasure is installed, everyone is happy, and nobody checks. Six weeks later the old habit returns. Closed is not solved. The work is done when the effect is verified, on a date you wrote down.

Our recommendations

  • Solve the problem before the form. A tidy page is not thinking.
  • Keep one metric. Use the same measure in the current condition, the target, and the follow-up.
  • Match every countermeasure to a cause. If an action has no cause, ask why you are doing it.
  • Draw, do not only write. One sketch, one chart, one photo beat a paragraph.
  • Print it and pin it up. An A3 on the board invites comments. An A3 in a folder does not.
  • Use A3 for learning, 8D for the customer. Many plants use both, for different readers.

Try it

Our A3 tool is coming soon. It will run in your browser, with no account, and your data will stay on your device. Meanwhile, see the tool page: /tools/a3.

From the book

This guide follows Chapter 6, the A3 chapter, of the field manual. See /book for the book.