What do you do if your Six Sigma manufacturing operations encounter a problem?

In my experience as CLSSBB practitioner, it is best to always know what the customer wants (i.e. Service Level Agreement or SLAs) and use it as your Ys and from there come up with your Xs or Key Performance Indicators (KPIs) to measure the success for each Y. Periodic meeting is also a great venue to discuss the RAID (Risk, Action, Issue and Decision) and you must always show the status, target dates and RACI (Responsible, Accountable, Consulted, Informed) for each of the items raised in your RAID log.

DMAIC: Define your Problem Statement in Measurable Terms. Favorite tool is the 5-Why Root Cause Analysis and never accept anything as “Operator Error”.

The question is very open. the first point when you find a problem is understand the "size of problem", the "size" is combination of impact and the frequency/quantity. this define the speed to solve. in other words, a very little issue with many quantities, has the priority than a critical issue with feel quantities. Other approach it is new or old problem, If it is old ( means already happen), check with supplier the solutions used and re-apply it or use a new solution because the solution already used isn′t good. ( use all 6sigma tools on it) to assure that isn′t happen again. If it is a new problem again, uses all 6sigma toolbox to solve it.

Define the problem clearly in detail will enable the black belt and green belts to put their efforts in the right direction. It also helps the management to understand the problem's importance very clearly along with its scope.

When dealing with a manufacturing problem, regardless if you are operating at six sigma, the most important thing is the team willingness to get to the real root cause and just not do the process steps defined in the company format for problem resolution. Quality & manufacturing professionals do the best problem resolution when they act like police officers, look at every single issue and from different angles. Defining the problem accurately is the key, problem well defined is half the solution. The second important thing is the data collection plan and the correct usage of the science & art behind data collection. I don't want to go more into details of biasness & GR&R and all that stuff. data should be clean and pure to work for you.

I would add here, making sure you can trust your data. Challenging the measurement system that triggered the problem is in my experience a critical step to make sure the problem is real, in some cases the problem was related to the method to quantify the variable. Always make sure you can trust your data before fixing and fake problem.

The measurement phase is a very important phase because having the correct data will help show the cost that the problem is having in your current operation. Don't be surprised if the way data is collected currently in your organization is not adequate and that some changes may need to be made. Data such as Downtime, Scrap cost, running at reduced rates, employee morale, etc. is all data that should be collected. One should also look at real time monitoring of data if your company has it available or look at implementing it. One of the most important things is to spend time on the floor observing the process. Look at the part/defect (talking to the part as some would say) will help to understand the issue along with talking to operator

Many times MEASURE phase takes a lot of time for the BBs and GBs. But smart BBs can set the measure phase with the right data collection plan that can reduce the efforts of data collection and still have enough data for analysis.

I have observed that the selection of right tools for analysis is just not enough. The data has to be analysed in interaction with other factors to identify the potential of the causes and root cause. Good analysis gives a great data for IMPROVE phase and vice-versa. It is important to analyse the data to the required depth with required number of data points.

I agree with this approach. Having said that, I would like to add that sometimes , data might be not be available or what ever is available, is not reliable. In such scenarios you may have to collect or do some anecdotal data gathering. What helped me? I went to the shop floor and did MUDA walk. Multiple times a day and multiple days. You will be surprised by the amount of information and insight you can gather. This definitely is a very strong tool in our arsenal.

Gather SME input on the problem and whether or not this problem has occurred in the past. Learn from previous solutions before trying to develop new ones.

I believe that when the operation is qualified as 6 sigma, there should be no issues. But if this happens, the entire system needs to be reviewed because if something goes wrong, we cannot guarantee that the entire process will be 100%

This is an important phase and how well the problem was defined and your measurement system is set up to collect data will contribute to the success of this phase. Look at different ways of slicing and dicing the data. If you have observed the process and have observations some cycles you should be able to see if the data matches. Use 5Y and Ishikawa fishbone diagrams along with the other 7 Quality tools. Look for trends and the differences between shifts and how things change over the course of a day. Use Pareto and focus on high volume items or the most frequently occurring issues. Look for abnormalities in the data to understand what is going on. Identify parameters that influence changes and determine high and low values to test.

From the analysis phase, one should have a good idea of parameters that will influence the output. Develop a plan to test these. Use the baseline data established in the measurement phase so that results can be compared. If modifying a process or making changes to equipment, use the baseline data and then what you think the changes will result in terms of scrap reduction, downtime reduction, or productivity increases and use these to justify the change in the process or equipment. Try to collect real time data if possible along with the collect of the critical parameters so that changes can be seen. Good standard work with examples along with training will help define the new solution. Audit the training to see if it is effective.

Control can be accomplished a number of ways. Implement new quality checks. Develop check sheets or design the process so that there is only one way to do it. Install sensors to identify if a process is going out of control. Implement control charts. From analysis and implementation phase, we should know the critical parameters. Develop control charts to monitor the critical parameters and ensure the process is in control. Implement periodic checks.

Schedule periodic audits to review the process and ensure that the new process is being followed. Look at installing sensors that can be used to alert operators if a given parameter is out of control. Go to the area where the change was implemented and observe the process to see if the process is operating the way you expected it to.,

First thing I do is go to the shop floor and talk to the employees… find out what the real problem is. I guess you could call that the “D” in DMAIC… stats and data may mean nothing if it’s a morale issue. Rule out the issue being the employees morale, overworking, or lack of their input before looking at numbers….

It depends on the type and complexity of the problem. Also the phase of Six Sigma, the problem is seen also is important - out if 5 phases D, M, A, I and C Resolution for each phase of Six Sigma journey differs. D phase needs redefining / course correction to the problem statements itself. M phase makes us to revisit the measure defined earlier and redefine ir device a new metric. A phase makes us to check the correctness in selection if analysis techniques and adopt more appropriate technique. I phase reflects the pace at which the improvement is getting done and compare it with the pace defined by Organization. C phase details about the laxity in controls that are defined and selection of better control mechanism.

It is important to get before and after pictures of the process and situation. Get people that are involved with the process involved. Try to determine if the issue is being caused upstream and the impact that the issue is having downstream. Often times a product that is having issues such as a print defect or a coating defect will cause not only issues with its operation but downstream operations as well resulting in poor first pass yield.