Navigating the Unknown – Identifying and Controlling Unknown Factors in Manufacturing

Engineer checking and maintenance technical data of system equipment condenser Water pump and piping air compressor system at manufacturing factory.

In manufacturing, managing both known and unknown factors is essential for stability and productivity. While known factors are controlled, unknown factors can cause unexpected disruptions. Detecting these with control charts and addressing them through systematic experimentation, like Design of Experiments (DOE), helps convert unknowns into controllable factors. This leads to more stable and efficient processes, improving overall performance.

Defining known factors and unknown factors 

There are two types of factors in every process: known factors, and unknown factors. Known factors are understood by the process operators and the known, strong impact ones are under control. Many of the factors in all processes are unknown but only some of them might have a strong impact on the process outputs (1). An example of known and unknown factors is presented below. 

 Known and unknown factors

Due to a lack of resources, it is impossible to determine every factor which has an impact on a process. As such, there will always be unknown factors influencing the process. 

Unknown factors and special cause variation 

Unknown factors are frequently the reason behind special cause variation. When an uncontrolled and unknown factor, especially a strong one, starts to influence the process, the variation of the process starts to increase, and the process becomes unstable. Control charts can notice special cause variation of different sizes. There are different rules for different violations, but the purpose of the rules is to help identify when the violation started. To explain this further, we will explore the basic violation according to Western Electric Rules (2). The rules are based on limits based on how many standard deviations the value differs from the mean. 

Rule 1

Rule 1 occurs when even a single data point falls outside the 3σ-limits, e.g. the value is over three standard deviations away from the centreline. This could be due to a strong unknown factor having an impact on the process. An example of this is presented in the graph below. 

Western Electric Rule 1

Rule 2 

Rule 2 occurs when two out of three values are outside the 2σ-limit on the same side of the centreline. This could still be a strong factor influencing the process, or even a medium strength factor deviating strongly. An example of this is presented below. 

Western Electric Rule 2

Rule 3 

Rule 3 occurs when four out of five values are outside the 1σ-limit on the same side of the centreline. This could be due to a weak factor being strongly deviated or a strong factor being mildly deviated. An example of this is presented below. 

Western Electric Rule 3

Rule 4

The final rule 4 occurs when eight consecutive points fall on the same side of the centreline. This indicates a mild impact on the process and even a weaker factor could be behind the variation. An example of this is presented below. 

Western Electric Rule 4


From unknown factors to known factors 

To fully tap into productive potential, it is necessary to convert as many unknown factors into known factors as possible. This has two benefits: firstly, it removes unnecessary special cause variation from the system. Secondly, by better understanding the system, one will likely increase the process yield through lower variation. Overall, the process becomes more stable and results in a higher yield. 

To change unknown factors into known factors experimentation is required. This is where Design of Experiments shines as a methodology (3). In Design of Experiments, factors are systematically modified to examine what kinds of impact they will have on the process. If the factor has a low impact on the process, it can be disregarded from further analysis. However, if the factor has a strong impact, it should be controlled. The figure below shows how the veil of the unknown becomes smaller as a new, strong factor is identified. 

The unknown becomes the known

Case example with the FH Cycle module 

Factory Harmonizer is an AI-based software solution designed for operators and engineers. Factory Harmonizer brings together the best of human expertise, machinery, and automated machine learning to stabilize the production process for better productivity. 

Factory Harmonizer’s newest module, FH Cycle, provides users with process condition values for batches, sequences, and measurements. The process condition index approximates batch quality. If a batch goes well, the process condition index will be of a high value while a bad batch will have a low value. 

Assuming we have a process, which is stable and under statistical process control, the condition index should fluctuate between values based on the current process capability. The value does not have to be 100 all the time to be stable, it only needs to fluctuate steadily between some values. This indicates that no major unknown factor is causing any significant disturbances in the process. 

If the process starts to degrade systematically, we should see this in the process condition index. As was discussed in a previous section, strong factors tend to show up faster than weaker factors in control charts. The control chart will give guidance on when to look for changes in the underlying process. In the example below there is a clear violation of rule 1 at the latest value. 

Example rule 1 violation

Identifying what might be causing production issues often relies on human intuition. Once potential causes are identified, experiments can start.  An example of a Design of Experiments with FH Cycle can be read in our previous blogpost. The purpose of the experiments should be to get the unknown factor under control. Assuming the experimentation is a success, the process should stabilize and return to normal as shown below. 

Stable production process


Understanding and managing both known and unknown factors is crucial for maintaining process stability and improving productivity. Unknown factors often introduce variability that can be detected through control charts and addressed using methodologies such as Design of Experiments (DOE). By systematically identifying and controlling these factors, processes can achieve higher yields and greater consistency.  

Want to hear more about how we at SimAnalytics help factories improve their productivity? Book a call with one of our experts.   



  1. Wheeler, D. (2008) Twenty things you need to know, SPC Press 
  1. Breyfogle III, Forrest W. (2003) Implementing Six Sigma: Smarter Solutions Using Statistical Methods, John Wiley & Sons; 2nd edition  


Image of the author

About the author

Lauri Lättilä
Head of Research, co-founder

Lauri is the true epitome of combining human and computational intelligence. There’s not a topic that Lauri hasn’t read about. He has been widely recognized from his work creating simulation models throughout industries. Loves his board games so feel free to challenge.

Related articles

Chemical lab
12/6/2024 Blog

Unlocking Process Efficiency: The Power of Design of Experiments in Manufacturing

Read article
Chemical factory
18/4/2024 Blog

A Short Guide to Common and Special Cause Variation in the Process Industry

Read article
Engineer on site in the sunset
6/2/2024 Blog

Factory Harmonizer as a tool for continuous improvement

Read article

Our Experts are here to help!

Heikki Juhola
Area Sales Executive

Book a demo