Kinesis CEM, LLC
Not All Customer Experience Variation is Equal: Use Control Charts to Identify Actual Changes in the Customer Experience
https://blog.kinesis-cem.com/2015/04/09/not-all-customer-experience-variation-is-equal-use-control-charts-to-identify-actual-changes-in-the-customer-experience/
Eric Larse is co-founder of Seattle-based Kinesis, which helps companies plan and execute their customer experience strategies. Mr. Larse can be reached at [email protected].
http://www.kinesis-cem.com
kinesis-cem.com 206.285.2900 [email protected]
Not All Customer Experience Variation is Equal:
Use Control Charts to Identify ActualChanges in the Customer Experience
Be it….
a survey of customers…
Variability is Common
mystery shops…
Variability is Common
social listening…
Variability is Common
or…
other customer experience
measurement…
Variability is Common
a certain amount of random variation in the data is normal.
Variability is Common
Managers need a means of
interpreting variation to evaluate if the
customer experience is truly changing, or
if the variation is simply random.
The Problem
Control Charts
One Solution
Statistical Tool to Measure
Variation Control Charts
Consider this example of mystery shop results:
For Example
May June July Aug Sept Oct Nov75%
77%
79%
81%
83%
85%
87%
89% Average Mystery Shop Score
Was July a bad month, or a random variation?
Does the improved performance in October and November represent a true change in the customer experience?
QuestionsMay June July Aug Sept Oct Nov75%
77%
79%
81%
83%
85%
87%
89% Average Mystery Shop Score
Control Limits
To answer these questions we need to set upper control
limits (UCL) and lower control (LCL)
limits.
Control Limits
Where:x = Grand Mean of the scoren = Mean sample sizeSD = Mean standard deviation
The following equations set control limits with a 95% confidence:
Control Limits
Applying these limits to our hypothetical mystery shop scores,
produces the following UCL & LCLs:
May June July Aug Sept Oct Nov75%
77%
79%
81%
83%
85%
87%
89% Average Mystery Shop Score
Questions
Now, we can answer our questions:
May June July Aug Sept Oct Nov75%
77%
79%
81%
83%
85%
87%
89% Average Mystery Shop Score
July was, in fact, a bad month.
While October is still with the threshold of random variation, November showed significant improvement.
Control charts give managers a meaningful way to determine if any
variation in their customer experience measurement
reflects an actual change in the experience as opposed
to random variation or chance.
Control Charts