What is Baseline Performance?

Learn about Baseline Performance and how to use this tool in Kure

Ken avatar
Written by Ken
Updated over a week ago

The baseline is a measurement of the current state of your process performance. You’ll compare the baseline to the desired customer requirement to help you view the process output from the customer’s perspective.

Ready to collect Baseline Data and start improving?

The Baseline should be shown in a run chart, plotting the performance of your Goal Statement measure over time. Later in your project, you’ll compare the improved process performance to the baseline to quantify the improvement gained.

You analyze the data to establish the current performance level, the baseline, and to look for two general types of clues:

  1. Baseline data pattern: look for non-random patterns. The baseline is displayed and analyzed using a Run Chart.

  2. Stratification: significant performance differences due to the levels of the stratification factors

Don’t overlook stratification factors! They can be key to finding some of the most important clues. Often your process will perform differently under various conditions. Think about who/what/when/where. If your process is done in different locations, that’s a great stratification factor. So is the day of the week, or sometimes weekend vs weekday. Teams or individual people can be good stratification factors. Remember we don’t want to blame people, but if there is a difference in performance with different people, we’d like to understand why.

How does Kure find Clues?

Kure makes all this easy by creating a data input form into which you can type your data directly, or upload it from a file. Kure automatically plots the Baseline chart for you and then asks if you see any non-random patterns.

This is important because we expect the baseline to simply randomly plot around its average ⁠— if it doesn’t, that be due to a root cause influencing it ⁠— and that can be a great clue!

Non-random patterns might include extremely high or low values, trends, or oscillation around the mean. We expect some “bouncing around” the mean, but in a random manner ⁠— if it is too “orderly”, that hints at a clue. You will be asked if you see any non-random patterns ⁠— if you do, you can enter a short description as a clue.

After this, Kure will look for non-random patterns and advise you of what it finds, offering the opportunity to log each as a clue.

Finally, Kure will test for performance differences based on the stratification factors by comparing the average performance for different values of each stratification factor, advising you of the result and offering to log as a clue.

In just a few steps, Kure does a considerable amount of analysis for you, so make sure you provide plenty of stratification factors so Kure has enough to help you out.

You can identify stratification factors in your Data Collection Plan.

Patterns Discovered

Kure expects to see random variation in your run chart. This means that the data points will be randomly distributed around the mean of your run chart. Kure analyzes your run chart data looking for evidence of patterns that are unlikely to occur if only random variation is present. Kure searchers your data for the following evidence.

Outlier Discovered

An outlier is a data value that is an extreme distance over or under the mean of the run chart. You could have multiple outliers in our data. In any sample of data you would expect to see one or a few data points that are the highest and lowest values. These are not necessarily extreme values. What Kure does is identify those values that are so high or low they are likely not random.

An extreme value indicates something unusual may have occurred in the process at that point in time to drive your performance level very high or very low. Examples could be a sudden increase in work volume, temporary staffing shortage, poor quality inputs, or an unusually difficult case being processed.

Explore the context and timeframe when extreme value occurred. Talk with the people working in the process during the time of the extreme value. Because Kure has identified it as an Outlier there is a high likelihood you can determine the reason. The reason is a clue you will explore later to identify potential root causes.

Trend Discovered

A trend is the movement of data points in one direction, increasing or decreasing. When a trend occurs on the run chart it is evidence that over a period of time something is steadily changing in the process. Possible reasons include steadily increasing or decreasing work volume, backlogs building in the process or staff learning curve.

If Kure detects a trend, you should investigate the process for the time period covered by the trend to determine the contributing factors. Be sure to involve those that are closest to the process and present during that time frame. The factors you identify are clues to help you identify potential root causes later in the project.

Oscillation Discovered

An oscillation is a repeating increasing and decreasing pattern in the data. While we expect there to be random variation in the run chart we do not expect to see a repeating fluctuation in the data. The oscillation indicates that there are factors in the process changing in a repeatable manner over time. When looking at data over the several months oscillations can occur due to seasonal factors. In the short term oscillation can be caused by factors that change by day of the week, hour of the day, week-to-week or shift.

When Kure detects an oscillation in your data, investigate the timing associated with the repeating pattern to determine your clue. Again, those closest to the process are likely to have valuable insight.

Stratification Difference Discovered

Stratification is sorting your overall data set into smaller data sets based on different conditions, or factors, that could influence process performance. Examples of such factors include day of the week, shift, location, type of order, etc. — anything that makes sense. Often, those with process knowledge can easily identify the stratification factors that likely influence process performance. You identified these stratification factors in the Data Collection Plan and included them in the data you collected.

Kure uses the stratification factors to sort your data into separate “buckets” then compare process performance of those buckets to one another. Kure identifies a Stratification Difference when the performance between the factors varies more than what would be expected with random variation. These Stratification Differences are clues to explore later in the project as you work to identify root causes and develop solutions.

Kure guides you through each step in collecting and analyzing Baseline Data by asking simple questions and providing guidance along the way. Powered by our Process Optimization Path® (artificial intelligence), Kure will help you and your teams collaborate to complete process improvement projects together.

Ready to collect Baseline Data and start improving?

Did this answer your question?