1. Not digging deep enough into the problem.
This manifest itself as either getting stuck on the physical causes which leads to replacing a lot of parts and solving the symptoms but not the true problem or even worse the human cause where it becomes an exercise in blame analysis. A gentleman by the name of Deming addressed this when he said "Blame the system not the people" The system or systemic cause allows for many if not all of the human mistakes to exist. If you want to learn more about the levels of root cause then check out this post.
2. Too many root cause investigations per month because the process triggers that kick off an investigation are too low.
The triggers can be set up to work off of a certain amount of downtime or cost or lost production or safety etc. or a combination of multiple elements but they should be set up to change as you mature into RCA. If you have the triggers set too low or if you are relying on management to ask for the investigation then you could have too many RCAs investigations and reports to complete each month. It is also important not to forget that for every RCA there is multiple action items that must be assigned, completed and verified. Can your systems and resources support that if you are busy generating RCA reports every time the wind blows.
3. Too much time spent on the report and not enough time spent on the implementation and follow up.
I am sure very few of you work for company that make RCA reports as a product. My guess is your company probably makes wigits and walumps and in the end a three pound final report does not add anything to the bottom line or any more margin on your widgets. Please remember the company is not paid by the pound of report, we are paid by the solution and the return on investment we generates.I have a one page report that I will share with you as a model for your reports if you send me an email
4. Lack of solid well understood and applicable tools and process.
Many sites have one side of this coin or the other and it is a major distraction. One site I visited in Brazil had great processes for problem identification and prioritization but the reliability engineers were using five whys as their only tool and were really missing the returns that they could have had. Another site had a great selection of tools to help understand the causes of the problem but their "corporate" RCA process had not been accepted and implemented. This lead to great findings and solutions but no way to ensure that they were ever implemented or if they truly solved the problem.
5. Don't ask the right question and you will miss major causal chains.
To help to mitigate this one use some thing like design and application review and follow up with change analysis to perfect your questions before you start your in-depth investigation and interviews. If you do this correctly you reduce the amount of time you will spend stopping and gather additional data during the analysis phase.
6. Findings and solutions are not verified to ensure that they effectively eliminated the original problem.
When this step is done correctly the site uses quantifiable metrics to measure the solutions that were identified and they are done at intervals that verify a sustainable long term solution.
7. Neglect to put focus on ensuring the best return on investment.
If you use the correct tools and you dig down into the problem you should be left with multiple possibilities for resolution of the problem. Once you have this identified you can evaluate the cost to implement and the effectiveness of the solution to see which action or combination of actions gives you the best overall return on the problem. In the end we change our thinking on analysis and it becomes less about finding "root cause" and more about finding the most effective and lowest cost mitigation or elimination strategy.
After reading through these seven common causes of "root cause failure" I hope you recognized a few and can eliminate them from taking the profits out of your process.
If you have question about any of these please feel free to send me an email.