When creating your project charter, you determined the Impacted Metric that you will use to measure your improvement, progress and success. It is critical to establish a baseline from which to measure improvement and to determine if or when you’ve met your objective.
In this lesson, we will provide a high-level overview of data analysis.
What data do you need? What data do you have? How are you going to collect it?
If you have no data or limited data to help you solve your problem, you need to develop a data collection strategy.
If you are not able to collect data systematically (via software or automatic data capture), you can collect data via:
You will need to document how and when your data was collected, so you can repeat the process to collect data during and after the improvement has been made.
If you are collecting data systematically through a software system or automated process, it is your obligation to fully understand the data collection process.
Often in continuous improvement, we are asking another department, process owner, data analyst or software owner to provide us with data.
When asking for data, sit down with the analyst or person providing the data. Review your project charter with the analyst to ensure he or she understands your objective. Tell the analyst how you intend to use the data.
Here are a few questions to ask:
- How is the data collected and input into the system?
- How often is the data collected?
- Tell me about the data quality. How is data quality ensured?
- Have any changes been made to the system or how data is entered? When were changes made?
Make sure you fully understand the data before you use it to make decisions.
Common Data Pitfalls
We see many common data pitfalls in CI projects:
- Relying only on instinct and personal experience
- Not "knowing" the data and not fully understanding the data collection process
- Relying only on the "average" or "mean" without looking at variability and trends
- Not digging into the true root cause and only addressing symptoms
- Utilizing convenient data
- Not using graphical analysis to visually depict the data
- Utilizing a single data point to make a decision without analyzing the trends
Here's how to avoid them:
- Utilize data whenever possible, collect data when data doesn't exist
- Understand how your data is collected and analyzed
- Capture statistics beyond the average - look at the median, mode, minimum/maximum values and standard deviation
- Always address the problem at its root
- Take the time to collect data. Even a short-duration sample can help validate (or invalidate) your assumptions
- Use graphical analysis to visually display the data
- Analyze data trends over time
If you need help with data analysis, reach out to your company's data analysts to help. Or, post in our community and we'll help provide you with insights!
Meet with your team to discuss your project data.
Update your Project Plan with the tasks needed to collect or analyze data.