Q 654. An important input while working with control charts is to decide the frequency at which we need to monitor the process. For e.g. - we could monitor the process every minute, every hour, every day etc. How is this frequency decided? What happens if frequency is too high or too low?

 

Note for website visitors -

• This platform hosts two weekly questions, one on Tuesday and the other on Friday. 
• All previous questions can be found here: https://www.benchmarksixsigma.com/forum/lean-six-sigma-business-excellence-questions/. 
• To participate in the current question, please visit the forum homepage at https://www.benchmarksixsigma.com/forum/.
• The question will be open until Tuesday or Friday at 5 PM Indian Standard Time, depending on the launch day. 
• Responses will not be visible until they are reviewed, and only non-plagiarised answers with less than 5-10% plagiarism will be approved. 
• If you are unsure about plagiarism, please check your answer using a plagiarism checker tool such as https://smallseotools.com/plagiarism-checker/ before submitting. 
• All correct answers shall be published, and the top-rated answer will be displayed first. The author will receive an honorable mention in our Business Excellence dictionary at https://www.benchmarksixsigma.com/forum/business-excellence-dictionary-glossary/ along with the related term.

• Some people seem to be using AI platforms to find forum answers. This is a risky approach as AI responses are error prone as our questions are application-oriented (they are never straightforward). Have a look at this funny example - https://www.benchmarksixsigma.com/forum/topic/39458-using-ai-to-respond-to-forum-questions/

• We also use an AI content detector at https://crossplag.com/ai-content-detector/. Only answers with less than 15-20% AI-generated content will be approved.

Control chart is a very powerful tool to constantly monitor a process performance. By implementing them on either input or output metrics, management can monitor progress and check each time a data point is out of control.

 

There are many types of control chart, depending on whether the data is discrete or continuous. Some of the examples of control charts are X-bar, P chart, U chart, NP chart, etc.

 

The frequency to monitor the process data points depicted on a control chart depends entirely on the frequency at which the data is reported/ analysed/ received. 

 

Let's take an example of an invoice processing process. The Turnaround time to complete the processing of each invoice is at an average of 2 hours. Therefore, every 2 hours we should see an additional data point on the TAT control chart. In this case, it is advisable to monitor the process every 2 hours so any anomalies (out of control data point) can be instantly caught and analysed. If this monitoring time is too long, SLA will be missed. If it is too short, it would add no value in monitoring the process.

 

Another example would be at a call centre. The AHT to complete a call is defined at an SLA of 5 mins (including hold time, after call work, etc). For the associates to achieve this, it is determined that at any point, there shouldn't be more than 3 calls in queue. Plotting calls in queue every minute, in a control chart, will help us understand spikes and special/ common causes for high call volume and this should ideally be monitored every minute or two. In case incoming call volume is monitored every one hour, instead of one minute, there would be multiple instances of calls getting dropped and/or expected AHT being missed.

 

Therefore, the frequency to monitor a control chart is dependent entirely on the nature of business and expected SLAs.

The monitoring frequency is the time interval between two monitorings. The frequency of monitoring is essential for guaranteeing results. Let's try to see this from a more down-to-earth angle through the example of hospitalized patients. In each hospital, there is a patient monitoring protocol. Indeed, depending on the severity of the illness and the urgency of care, the frequency of monitoring will be more or less frequent. imagine that for patients admitted to intensive care, the monitoring frequency is further away and that for patients awaiting discharge, the monitoring frequency is closer. what is likely to happen? The operations will not be adapted to the needs and the risk of loss of human life will be quite high.
Given the previous example, we can say that the business process is comparable to hospital protocol. Therefore, deciding how often to monitor a process with control charts is crucial to the survival of a business. This will help detect and respond effectively to process variations. When determining the frequency of monitoring a process, it is important to consider factors such as the stability of the process, the type and quantity of product data, the costs and resources allocated to monitoring, the sensitivity of the process and the contribution of the process to the overall results of the company.
The stability of the process is almost the most determining factor in deciding the monitoring frequency. In fact, the more stable the process is, the less the process will be monitored. and of course, inversely proportionally. If the process is unstable, it will be necessary to systematically monitor it at regular intervals. This will allow variations in performance to be detected early and to more or less have time to correct through controlled actions. On the other hand, if the process is stable, the monitoring frequency will be much less.
The pace of data production can also be restrictive for process monitoring. Indeed, for certain cases where data is produced at a slow rate, it is often difficult to closely monitor the processes. If the process generates a large volume of data quickly, more frequent monitoring may be feasible and beneficial. Large-scale manufacturing, with the volume of data generated, must be monitored very closely. If necessary, in the event of a variation in quality, non-conformities will be considerable.
The decision to monitor more processes to the detriment of others may lie in the criticality of the latter. If a process has a major contribution to the quality of the product or service, it may justify more frequent monitoring to guarantee performance with very little variation from targets.
However, the cost of monitoring may be considerable and requires more resources in terms of time and investment in equipment, software and data storage devices. making it a factor to consider when determining monitoring frequency.

Because of the above, it is important to talk about the possible consequences of incorrectly sizing the process monitoring frequency.
If the monitoring frequency is too high, it can consume excessive resources. Indeed this will generate waste (MUDA) such as overproduction, overprocessing and an irrational use of resources thus leading to an increase in costs.
This frequency, more than necessary, makes it difficult to distinguish between normal variation which does not impact the final quality from those which negatively impact the latter. This leads to excessive adjustments to correct minor variations with unnecessary interventions that can disturb the stability of the process.
On the other hand, if the monitoring frequency is too low it can result in delayed detection of variations and deviations in processes, allowing problems to escalate before they are detected and addressed. Therefore, process owners could lose control and this will not militate in favour of the quality of the product or service and will certainly degrade CSAT and NPS. When it comes to continuous improvement, optimization opportunities will be missed. Indeed, certain root causes of problems can remain unsuspected for a fairly long period.
In conclusion, to guarantee the stability and control of a process, the frequency of monitoring is important. moreover, since factors such as stability, control, and resources consumed by a process are not static, it is essential to review the monitoring frequencies regularly to maintain a certain alignment between the performance of a process and the monitoring frequency of the latter. Therefore, the frequency of process monitoring can be revised upwards or downwards depending on the results obtained. This will continually improve the said process and overall business results.

Control charts are being used to study the variation in the process. Statistical control limits establish process capability.

Statistical control limits are another way to separate common-cause and special-cause variation. It separates the special causes from the common causes. Control charts can give early identification of special causes so that there can be timely resolution by the users, before poor-quality products are being made. To control a process using control charts, the monitoring should be of key process input variables, and the process flow is stopped for resolution when this variable goes out of control

 

Assume the scenario that most of the data points are out CL due to high frequency. Then it is next to impossible to control the process performance. In general a control chart frequency of any KPIs, is completely depends on the feedback or response time (Change adjustment) from the measurement and the measurement frequency.  

 

For Example, the control chart is tracking the porosity of the paper, for which the online measurement is giving data in every 5min data & lab measurement in every 30mins. Online gauge trend is follows the lab gauge. If any issue of porosity, the user adjust some input parameters to bring back into control, this process takes 10 mins. Hence based on the the above data, we consider 5mins is the frequency of control chart. The lab gauge value used to calibrate or validate the online gauge. So here i am concluding that frequency is always depends on the response of time, represent the source of variation & measurement frequency. It is believe that, if we cant measure the change of any KPI, then we cannot control.

 

Frequency become more crucial when the process is having more time lag.

 

If the users making the control chart for the out put parameters, it always recommended to keep the specification limits along with control limits. Some times if the process data showing out of control limits, but within specification limits, then user no need to worry, there is some to adjust back the process to keep the all the good produced are with in customer specification. 

Control chart monitoring frequency is determined based on various factors such as the type of process, stability of the process, sensitivity requirement, and other practical considerations related to data collection and analysis. Listed below are some considerations for determining the frequency of monitoring control charts:
Process Variation: If the process is stable and consistent over time and there are no significant changes expected in the process then monitoring with less frequency would suffice. The Opposite is also true, if there is high variability in the process or it is prone to frequent changes then the monitoring frequency required may be higher to detect trends promptly.
Process criticality: The impact of any process on product quality, and customer satisfaction will influence the control charts monitoring frequency. The higher the impact, the more frequent the monitoring, and vice versa.
Data accessibility: Availability of data also influences monitoring frequency. If data is readily accessible in real time then more frequent monitoring is possible. However, if data collection requires significant time or resources then less frequent monitoring may be a more practical option.
Statistical consideration: The desired level of sensitivity can also guide the selection of frequency monitoring.

For example: for narrower control limits or small sample sizes control charts will require more frequent monitoring and for wider control limits and high sample sizes less frequent monitoring will also suffice the requirement to maintain sensitivity to process change.
Operational bottlenecks: Various process constraints such as resource availability, and operational priorities also influence the frequency of monitoring which is why organizations balance the need for timely monitoring with resource availability and competing priorities.
Statutory requirement: In Industries that are highly regulated such as Pharma, Chemical, Oil 7 Gas, etc. there are statutory requirement or standards that prescribes specific monitoring frequencies for certain processes and quality metrics. In order to comply with these requirements organization ensures the frequency of monitoring as per the prescribed specification.
Data analysis: By analyzing historical trends or process data could provide insights to derive appropriate frequency for monitoring control charts. Various patterns of variability, seasonality, or past process changes can guide in deciding optimum monitoring intervals.

Frequency levels in control charts are very critical aspects as they enable the data engineer to understand process variability, maintain quality standards, and provide insights to select appropriate corrective action.
There are certain repercussions associated with the wrong consideration of monitoring frequency level, such as:

1. If the Frequency level is too high then monitoring control charts would require excessive resources in terms of time required, no. of people deployed, and efforts required to collect data. 
Frequent monitoring can also lead to an increase in the chances of detecting random variability in the process also known as noise. It can also lead to a decrease in the control chart's sensitivity to detect impactful process variations. With continuous monitoring, it becomes difficult to differentiate between process shifts causing process instability. Continuous or frequent monitoring can also lead to information overload and if the collected data does not make any sense to the decision-makers then they tend to lose focus on key performance indicators and priorities.

2. Low Frequency monitoring level poses the risk of delayed detection of process variability or abnormalities, There will exist the possibility that significant process changes may go unnoticed and could cause issue escalation resulting in the requirement of higher resources to perform corrective action.
Decision makers could also miss the opportunities to check and address process inefficiencies, defects, or quality issues this could lead to the suffering of improvement initiatives due to lack of timely feedback.
Because of reduced monitoring frequency decision makers can also experience a risk of slow response time to process variability.
Eg. If the customer is not satisfied due to any particular process then having a lower frequency interval can lead to an increase in time to detect the process variation cause of customer dissatisfaction and can increase the time span till which customer remains dissatisfied and can also impact brand image.

Great insights in the published answers. Recommend readers to review all the answers.

 

The winning answer has been written by Ousmane Fall. Well done!