Third - order logic of process cycle: A systematic methodology from analysis to improvement
1. Process analysis: Use data to clarify the gaps among "objectives - current situation - risks"
Process analysis is the "starting point" of the entire cycle. It is not a "perfunctory survey," but rather uses quantifiable standards to transform the vague "process" into "measurable problems." The core is to answer three questions:
1.1 Define the core objective of the process: Clarify "what should be done"
First, "set the tone" for the process: What are its inputs (raw materials, data, customer requirements)? What rigid requirements must the output meet (product specifications, service timeliness, compliance standards)? Where are the boundaries (which steps belong to this process and which do not)?
- For example, the core objective of an automobile engine assembly line is to use qualified components (input), complete 120 assembly steps (process) in accordance with the SOP, and produce engines that meet the requirements of "fuel consumption per 100 kilometers ≤ 6L and failure rate ≤ 0.1%" (output).
- Another example: The core goal of the order - dispatching process for food delivery riders is to receive user orders (input), match the nearest rider within 30 seconds (timeliness), and ensure that the food is delivered within 30 minutes (output).
The key to this step is to transform "vague requirements" into "verifiable standards". Without clear goals, subsequent control and improvement will lose their direction.
1.2 Identify the risk points in the process: Anticipate "what might go wrong"
The fluctuation of the process is inevitable, but risk is "predictable fluctuation". Three types of problems should be analyzed emphatically:
Sources of variation: What factors can cause the process to deviate from the target? For example, materials (fluctuations in the hardness of steel), equipment (wear of machine tool cutters), personnel (operational errors of newbies), and environment (the humidity in the printing workshop affects the expansion and contraction of paper) — these are all common sources of variation.
Known variability: Determine the existing fluctuation range of the process through historical data. For example, the screw torque of a certain production line has fluctuated within ±5 N·m in the past 6 months, or the defect rate of a certain product has always remained between 0.5% and 1%.
Sensitive parameters: Which variables will cause major problems with just a slight change? For example, the temperature in a chemical reaction (a difference of 2°C will produce by - products), and the "sugar amount setting" in a milk tea shop (an extra 5g will make it too sweet). These are the "weak points" of the process. Once out of control, they will lead to batch scrap.
This step is to "detect mines" in advance – knowing where problems are likely to occur so that subsequent control can conduct "precision strikes".
1.3 Evaluate the current performance of the process: See clearly "what is actually being done"
Use data to verify whether the process is truly achieving the goals, and focus on four dimensions:
Waste situation: Is there any scrapping (defective products are directly discarded) or rework (mobile phone cases with poor injection molding need to be re-polished)? For example, the rework rate of an electronics factory reaches 8%, and it incurs an additional cost of 5 million yuan each year—these are all visible losses.
Statistical stability: Is the process in a "state of statistical control" (i.e., only subject to variation due to common causes and without special fluctuations)? It can be judged by control charts. For example, there are no points outside the control limits on the X - R chart, or there is no non - random pattern (such as 5 consecutive points rising) on the P chart (defect rate chart).
Process capability: Can it meet the specification requirements? It is measured by CPK (Process Capability Index). For example, if the product specification is ±0.5mm and the process standard deviation is 0.1mm, CPK = 1.67 (excellent). If the standard deviation is 0.3mm, CPK = 0.67 (insufficient, improvement is needed).
Reliability: Is the process continuously stable? For example, the equipment downtime rate (a certain production line is shut down for 10 hours per month, affecting 2% of the output), the number of process interruptions (the express sorting line jams 5 times a week, delaying 100 orders) - these will directly affect efficiency and customer experience.
This step is to "look at oneself in the mirror" with data - to clarify the "health level" of the current process and avoid misjudgment of "feeling good about oneself".
2. Process control: Maintain "stability in dynamics" through monitoring
The process is not a "static machine" but a "living system" - equipment will age, personnel will move, and material suppliers will change. Therefore, stability does not mean "remaining unchanged" but "correcting deviations in a timely manner".
The core of control is "prevention first": intervene before adverse changes occur by monitoring key parameters in real - time, rather than "putting out the fire after the fact". Common tools include:
Control chart: For example, the X-R chart monitors continuous data (such as product dimensions), and the P chart monitors the defect rate (such as the proportion of defective products). The production line measures 10 products per hour. If a point exceeds the control limit (e.g., the dimension exceeds ±0.3mm), stop the machine immediately for inspection (such as checking if the cutting tool is worn).
Real-time monitoring systems: For example, the DCS system (Distributed Control System) in chemical production, which tracks parameters such as temperature and pressure in real time and immediately gives an alarm once they exceed the standard; the WMS system in e-commerce warehouses, which monitors the picking timeliness. If there is a delay of more than 10 minutes in a certain area, it will automatically assign workers for support.
Standard Operating Procedure (SOP): By specifying details such as "shake the cup 10 times and seal it for 1 second", it reduces differences among personnel. For example, the SOP in a milk tea shop can ensure that the sweetness of each cup of milk tea is consistent, avoiding fluctuations like "customers complaining about excessive sweetness".
The goal of control is to "lock" the process within the "qualified range" - to avoid quality deterioration or efficiency loss caused by fluctuations.
3. Process improvement: Break through the "bottleneck after stabilization" with data
Control means "maintaining the status quo", while improvement means "reaching a higher level". When the process is already stable but still has room for improvement (e.g., insufficient CPK, high cost, low efficiency), a more systematic method is needed to optimize the variables.
The key to improvement is to "drive decisions with data" rather than make changes based on "gut feelings." Commonly used tools include:
Design of Experiments (DOE): By systematically changing variables, identify the key factors that affect the results. For example, if you want to improve the battery life, test three factors: "material (A/B), charging current (1A/2A), and temperature (25℃/35℃)". Use a full - factorial experiment to conclude that "the material is the key factor", and then replace it with a better material, resulting in a 20% increase in battery life.
Advanced control charts: For example, the EWMA (Exponentially Weighted Moving Average) chart is more sensitive than traditional control charts and is suitable for monitoring small fluctuations (such as the defect rate of semiconductor chips, where a 0.1% change can affect the yield rate); multivariate control charts are used to monitor multiple related parameters (such as temperature and pressure simultaneously affecting the reaction results).
Process Reengineering: Use "Value Stream Mapping (VSM)" to identify the "non - value - added processes" in the process (such as the "secondary handling" in the warehouse), and reduce the handling time by 30% through layout adjustment. This is more effective than "adding manpower".
The goal of improvement is to "break through the bottleneck" - to push the process from "qualified" to "excellent", or from "excellent" to "outstanding".
The essence of process cycle
Process analysis is about "understanding the current situation", process control is about "maintaining the bottom line", and process improvement is about "raising the upper limit". The three are closely linked, not a "one-time task", but a continuously iterative cycle.
For example:
- Through analysis, it is found that "the CPK of a certain production line is only 0.67 (insufficient)".
- Monitor the dimensions through the control using the X-R chart and keep the fluctuation within ±0.2mm.
- By using DOE for improvement, it was found that "tool material is the key factor". After replacement, the CPK increased to 1.33 (excellent).
- Go back to the analysis and see if there are any new sources of fluctuations (such as whether the lifespan of new cutting tools is shorter)..
This is the core of the process cycle: using data to drive continuous optimization and enabling the process to shift from passive response to active evolution.