Root cause analysis template
A root cause analysis (RCA) finds out why a problem happened, not just what went wrong, so that the fix stops it coming back. This free Word template takes you through it in order: the problem statement, a timeline, the possible causes and the evidence for each, the root cause, corrective and preventive actions, and a later check that the fix worked.
Running a company of 30 to 200 people? Documents like this are now worth money.
Use it when something has cost real money, hurt or nearly hurt someone, upset a customer, or keeps recurring however often it's patched. For a small, simple problem, the 5 whys may be all you need.
What's in the template
Root cause analysis
| Problem | One line, e.g. wrong items shipped on 14 orders |
|---|---|
| Reference | |
| Date it happened | |
| Date it was found | |
| Lead | |
| Team | People who do the work, plus someone independent |
| Sponsor | Approves the actions |
1. Problem statement
Facts and numbers only: no causes yet, and no names.
| Question | Answer |
|---|---|
| What went wrong, and where? | |
| When did it start, and when was it noticed? | |
| How big is it (orders, money, people, hours)? | |
| What should have happened instead? |
2. Containment
| Immediate action to protect customers, staff or money | By | Date |
|---|---|---|
3. Timeline
| Date and time | What happened | Evidence (record, log, witness) |
|---|---|---|
4. Possible causes
Generate them with a 5 whys chain or a fishbone diagram, then test each one against the evidence.
| Possible cause | Evidence for | Evidence against | Verdict |
|---|---|---|---|
5. Root cause
| Level | Cause | How we know |
|---|---|---|
| Immediate | ||
| Underlying | ||
| Root |
- Removing it would stop the problem recurring
- It's within our control to change
- It explains every confirmed fact in the timeline
6. Corrective and preventive actions
| Action | Corrective or preventive | Owner | Due |
|---|---|---|---|
7. Effectiveness check
| Measure | Target | Check on | Result |
|---|---|---|---|
8. Close-out
| Documents to update | SOPs, risk register, training records |
|---|---|
| Approved by | |
| Date closed |
How to run a root cause analysis
- Contain the problem first. Protect customers and staff straight away (hold the stock, correct the payments, fence off the area), and keep the evidence while you do it.
- Write the problem as a measurable fact. "14 orders shipped with the wrong frying pan between 3 and 7 March" can be tested. "Picking errors" can't, and "Sam picked the wrong items" has already decided the answer.
- Build the timeline from records such as system logs, emails, delivery notes and rotas, then talk to the people involved while memories are fresh. Make it clear this is fact-finding, not fault-finding.
- List every possible cause before picking one. A fishbone diagram gives you breadth; a 5 whys chain gives you depth on the strongest candidates.
- Test each cause against the evidence. Expect more than one root cause: OSHA's guidance says there often is.
- Choose actions that change the system. HSE's investigation workbook ranks fixes: remove the risk, tackle it at source, and only then rely on people following a procedure.
- Set a date to check the result, and close the analysis only when the problem has stayed away.
Immediate, underlying and root causes
Most analyses find causes at three levels. HSE uses these terms for investigating accidents, and they work just as well for quality and service problems:
| Level | What it is | Example |
|---|---|---|
| Immediate | The most obvious reason the event happened | The picker took the 24 cm pan from the bay |
| Underlying | The condition or practice that allowed it | Both pan sizes had shared one bay since the stock count |
| Root | The management or system failing the others grow from | Picking checks the bay's barcode, never the product's, so any storage error reaches customers |
Fixing the immediate cause only stops that exact sequence recurring. Fixing the root cause stops the whole family of problems it produces.
Example: wrong items shipped at a fictional wholesaler
An analysis by a fictional 70-person kitchenware wholesaler. Problem: between 3 and 7 March, 14 orders shipped with a 24 cm frying pan instead of the 28 cm one, found through customer complaints; returns, re-delivery and credits cost £1,900. Containment: every open order for either size checked and six re-picked; the bay emptied and relabelled.
| Possible cause | Evidence | Verdict |
|---|---|---|
| Pickers careless | Four different pickers made the same mistake; their error rate on other products was normal | Ruled out |
| Both sizes stored in one bay | Location history shows both in bay C14 from 1 March, after the stock count | Confirmed (underlying) |
| Scanner checks the bay, not the item | The handhelds confirm the location barcode only | Confirmed (root) |
| Look-alike supplier cartons | Both sizes arrive in plain brown cartons with small labels | Contributing |
Actions: make the handhelds require a product scan at every pick (corrective); stop the warehouse system accepting two products in one bay (preventive); ask the supplier to print the size large on each carton. Effectiveness check: wrong-item complaints across all pans, reviewed weekly for eight weeks, target zero.
Common mistakes
- Stopping at "human error". OSHA warns that concluding carelessness or a failure to follow procedure alone caused an incident misses the underlying causes. Ask why the error was possible and why nothing caught it.
- Putting the answer in the problem statement, such as "lack of training", so the analysis only confirms it.
- One manager doing it alone. The people who do the work know where the process really breaks.
- Actions that are all reminders and retraining, which leave the process that allowed the error unchanged.
- Closing it when the actions are done, rather than when the measure shows the problem has gone.
- A full RCA for everything. Save it for problems that are costly, serious or keep coming back.
Frequently asked questions
What's the difference between a root cause and a contributing factor?
Removing a root cause would stop the problem recurring. A contributing factor made it more likely or worse, but removing it alone wouldn't. In the example, the look-alike cartons contributed; the scanner set-up was a root cause.
Which root cause analysis tool should I use?
The 5 whys suits a fault with one obvious thread; a fishbone diagram suits a problem with many possible causes; a complex incident needs a timeline first. OSHA's root cause fact sheet also lists checklists, logic or event trees and sequence diagrams, and IEC 62740, the international standard on root cause analysis, sets out each technique's strengths and weaknesses.
What's the difference between corrective and preventive action?
A corrective action removes the cause of a problem that has happened, so it doesn't happen again. A preventive action fixes the same weakness elsewhere before it causes trouble there, such as every similar bay or product line.