The most well-run operations strive for continuous improvement, and what differentiates them from competitors is their ability to identify root causes of problems and implement solutions.
Quality management gurus have developed, identified, and promoted a number of investigative and analytical tools. In this article, we’ll cover the six most widely adopted RCA tools and how to use them, including:
The Pareto chart is a visual representation of the 80/20 rule. It’s the principle that states that 80% of outputs result from 20% of inputs. Though only an empirical observation, the 80/20 rule often applies to situations with randomness and large populations. In problem-solving, it’s used to identify the issues or topics making the biggest contribution to a problem.
Consider a situation where there’s waste due to quality problems. Defective products are sorted into categories. These are arranged in descending order, so the category with the most items appears first. In addition, the percentage of total products involved is calculated for each category.
Categories are then displayed as a bar chart. An additional line graph plots the cumulative number of products involved by category in descending order. The resulting chart is an excellent visualization of why it makes the most sense to address the biggest item first: because it will have the greatest impact.
Businesses can analyze machine downtime the same way. Using data on either time lost to specific causes or the number of stoppages, a Pareto chart can identify the biggest contributors to a problem.
See how Amper’s operational analytics can help you understand what shift, machines and/or operators are struggling with what issues.
Used extensively in lean manufacturing kaizen efforts, the five whys approaches the problem-solving process like peeling an onion. When using this technique to approach a problem, you ask “why?” five times. In theory, each time you ask “why?” (for example, “Why did that happen?” or “Why is this done a certain way?”) you reveal another piece of the greater problem.
For this RCA method to be effective, it’s best to direct your questions to people with detailed knowledge of the products, processes, or equipment. Each time you ask why, the knowledgeable individuals are pushed to think deeper about the reasons. In this way, a questioner who lacks expert knowledge can help those who do dig down to the root cause.
The five whys method is effective in many situations. For example, it can be used to determine the root causes of production downtime, or to understand what caused an accident.
The five whys is one of the easier problem-solving tools, but requires thought and active listening. It’s important to explain the goal of the exercise to the people being questioned before you start questioning them — otherwise there’s a risk they may feel patronized or insulted.
When investigating a problem like unplanned machine downtime, you may wish to follow a format like this:
Q: Why is the machine stopped?
A: Because we ran out of material.
Q: Why did you run out of material?
A: Because the forklift didn’t bring the next pallet.
Q: Why didn’t the forklift bring the pallet?
A: Because no one asked the driver.
Q: Why did no one ask the driver?
A: Because we couldn’t find him.
Q: Why couldn’t you find him?
A: Because he was busy on the other side of the factory.
In this situation, it appears the root cause is that there is no timely way of asking the forklift driver to bring the next pallet.
A scatter diagram is a tool for exploring cause and effect relationships. (It may also be called correlation analysis and be subjected to intensive mathematical treatment.)
Scatter diagrams are useful when input and output variables can be expressed numerically. They take the form of a chart with one variable along the x-axis and the other along the y-axis. Typically there is a hypothesis that the y values result from, or are in some way related to, the x values.
One application for scatter diagrams is in helping to understand how an input variable (such as material thickness) affects an aspect of the output (like radius after bending). If a relationship exists, the individual data points will appear as something close to a line.
Other examples can be seen in processes like casting or molding where it’s useful to determine if a relationship exists between input and output variables. If metal or plastic temperature is the input variable, the output could be solidification time or final weight.
The scatter diagram requires measurable input and output variables. Furthermore, it should be possible to vary the input value over a range in order to see if this changes the output.
Scatter diagrams must be interpreted with care. As is often noted, correlation between variables does not necessarily indicate causation — there may be other causal factors at work.
The fishbone diagram, sometimes called an Ishikawa diagram after its inventor, Dr. Kaoru Ishikawa, is a powerful tool for identifying possible causes of a problem. It gets its name from its resemblance to a fish skeleton.
When faced with a problem, the fishbone diagram is one of the first root cause analysis tools to use to help a team identify possible causes. Fishbone diagrams dissect problems by using six categories:
A fishbone or Ishikawa diagram is used when it’s unclear what could have caused a problem. The idea is to help investigators focus their attention. Typical scenarios would be:
To complete an Ishikawa diagram, present a problem statement to your team and ask them to brainstorm possible causes, using the six Ms as a guide. To be effective, the team should have some relevant knowledge, but they need not be experts. In fact, some authorities suggest experts may be too quick to jump to a potential cause before completing a thorough analysis.
Use Amper’s Fishbone Diagram template here.
FMEA is a proactive tool for problem prevention. In practice, it’s often used when designing a new machine or process as a way to determine what might go wrong — and therefore what to guard against. Another use is to help design longer-lasting, more reliable products.
FMEA begins by developing a list of things that might go wrong — these are the failure modes. These failure modes go into a template worksheet. Next, your team identifies the consequences of each possible failure mode along with estimates of probability and impact. From here the team focuses on identifying actions to prevent (or at least mitigate) the most serious failures.
Using an FMEA template, a team identifies possible failure modes. For each one, they assess severity, probability of occurrence, and probability of detection. Each is assigned a numerical value and multiplied to arrive at a risk priority number (RPN). The appropriate management team members can then use the RPN to prioritize remedial or preventive actions.
This top-down root cause analysis tool uses boolean logic (AND, NOT, and related operators) to determine the cause or causes of a problem. A strength of fault tree analysis is that it allows for specific combinations of events, and uses probabilities to quantify risks and likelihoods. This level of specificity makes fault tree analysis one of the more complex root cause analysis tools.
Fault tree analysis is good for investigating how a particular combination of events or conditions leads to a certain outcome. For example, it could be used to understand what happened leading up to a fire or industrial accident. Another use would be to discover how a faulty product passed through quality assurance checkpoints.
The goal of fault tree analysis is to identify the cut set — the specific combination of events that created a problem. Start by identifying the immediate contributors, then link these to the problem with logical operators. Next, identify second-level contributors and link these to the first level in the same way. Continue until you reach the root causes. Then, calculate the probabilities up through the tree to determine the likelihood of the problem happening (or happening again).
When manufacturers don’t perform root cause analysis, the same problems will keep happening. Leaders use appropriate root cause analysis tools to identify why problems occurred, and then implement corrective action to prevent recurrence.
Lean manufacturing experts, quality gurus, and others have developed many tools for performing root cause analysis. Of the six tools we’ve discussed above, only the five whys can be used without data surrounding the problem: The others all benefit from detailed production metrics,, machine uptime records, and similar data.
Historically, gathering manufacturing data has been difficult. Amper simplifies this with a range of tools developed to provide visibility into the most complex manufacturing operations. Learn more about what our production monitoring solutions can do for your facility — request your demo today.