Data-driven problem solving
What a glimpse of problem-solving in the industry using data and statistics looks like.
by Sharihan.NB
The problem of understanding the problems and the deep dive
Clearly state your problem
Problem-solving is a methodical approach to resolving problems, to close the discrepancy between current and desired states. Solution finding is the more positive way of putting it. In everyday life, the problem is something like gaining weight over the holiday season as a result of eating an excessive amount of chocolate chips. The problem in the industry is something related to a decrease in product yield after switching chemical suppliers. This post will explain ways to solve problems using a data-driven approach to decision-making.
The range of problems is broad. We can classify problems based on how difficult they are to solve, how badly they can affect the output, the provision of relevant data and ability, whether or not the solutions are executable/cost-driven, and how many resources we need.
In life, we can’t expect to have no problems at all; instead, we’re just getting better at dealing with problems in such a way that they don’t seem to be problems. How do we go about it? How do we be better in solving it? Most of what is said are based on one’s own or someone else’s personal experiences. If we dive deeper into the elements of experiences , we will find plenty of information made up of so many data.
The data-driven approach is the highly recommended method for defining, deciding and solving the problem. Asking the right questions will help to clearly define the problem statements. Refer to four questions below.
- What is the problem?: Low Yield
- When it happened?: Daily
- Where it happened ?: Print Process
- How much is the impact?: 90% to 68%, Scrap Loss cost $500,000, turn-around set-back dropped to below 60%
Set expected outcomes
The project objectives will specify how much progress is expected and when it will occur. This is done for a variety of reasons, including having an agreeable target, coordinating team efforts, and ensuring stakeholder alignment. The goal, like problem statements, can be stated clearly; it includes the following key elements:
- What: Characteristic of that will be measured
- When: the time frame from current state to desired future state
- Where: at the problematic process which is the print process
- How much: the values of improvement
Process knowledge
Tools for better understanding the process
A process is a series of actions that are carried out in order to achieve a specific result. In general, processes take inputs and provide outputs. You can use the process mapping technique to gain a thorough understanding of a process. Value Stream Mapping is one of the tools that is widely used when it comes to improving fab efficiency or eliminating non-value-added activities in the process. There are other types of maps with different purposes such as SIPOC, High-level Process Map, or Detailed Flowchart. Remember that it is important to have the right experts in the team, the machine’s owner and others involved in the process, get feedback from upstream and downstream associates, go through each step one by one, study the manuals and finalize the Process Map with other team members.
Generate Ideas and Identify Causes
Tools for generating ideas and identifying potential causes
To generate ideas and gain process knowledge, key technical
teams rely on subject matter experts and creative ideas. This activity is called
brainstorming where everyone will contribute their no-overthinking ideas. When
combined with the Root Cause Analysis (RCA) methods provided below, these keys
will aid in identifying potential and root causes.
- Brainstorming
- Pareto Chart
- Fishbone Diagram
- 5 Whys
- Failure Mode Effects Analysis (FMEA)
Actions. Improvement and post-implementation
The most effective problem-solving strategy
Making data-driven decisions is central to the statistical method. The argument for data is that data does not lie. All processes have the ability to generate information, and they are all connected. Every process has variations; understanding and reducing variation is absolutely key.
To understand the variations, you’ll need to do some statistical thinking. This involves data collection (historical data, current data, test data), plotting data to the graph for graphical aids in extracting the information from trends such as maximum and minimum values, distributions, statistical tests, and so on. Then, this knowledge will influence your decision-making, you’ll want to understand important things like the variation’s characteristics, whether it’s a random variation, or if real changes have occurred.
The use of control charts will point you in the right direction. You don’t want to have to adjust your procedure every time your parameters deviate slightly from the norm. It is time-consuming and costly. It is the inverse of efficiency.
Once the root cause of a problem has been identified and potential solutions have been identified, strategic execution is critical and must be monitored and evaluated over time. It is critical to define the timeline and assess overall performance following implementation in order to officially close the problem as agreed upon by your team members. You now have solid data on a solved problem to which you can refer in the future.
Contact our HUJANPERA team for professional data-driven statistical services. Visit our projects page for further example on completed jobs.
Source: https://www.jmp.com/en_my/online-statistics-course.html