How do you describe this theory? Customer complains decrease over time as Total Customers volume increase is incorrect to look at total volume.

#1
Hi everyone

I would like to know how do you describe the below referring to some statistical theory.

In Yr 1 - when the customers base in 1000 there is 100 complaints that is 10% of total.
Over time the customers growth is a curve shape and by Yr 9 reached its peak.
In Yr10 - when the customers base has increased over time to 1 Million (Only 10,000 customers added between Yr9 and Yr10) and the complaints is 15,000. Which is 1.5% of the total.
We seem to be doing a better job managing complaints now however that is untrue. Of the 15,000 complaints 2500 complaints relate to new customers added between Yr9 and Yr10 while the rest belong to older customers. So I know here we should look at other factors to determine the logic like group the number of years the customer with company and then look at the trend and determine if we are doing a better job thus reduction in complaints.
My question is what theory should I refer to advise we should not report on total customer base to determine if we are doing well or not.
 

AngleWyrm

Active Member
#2
A proportion of 100/1000 complaints/customer that changed to 15k/1M complaints/customer over time appears to have a significant distinction of newness of customer. How can we better see that?

Newness of customer changes over time, but it could instead be represented by a first contact date, which doesn't change over time. Then each year's complaints/customer could be shown as a stacked bar chart, with each year's bar subdivided into portions by first contact date.
 
Last edited:
#3
A proportion of 100/1000 complaints/customer that changed to 15k/1M complaints/customer over time appears to have a significant distinction of newness of customer. How can we better see that?

Newness of customer changes over time, but it could instead be represented by a first contact date, which doesn't change over time. Then each year's complaints/customer could be shown as a stacked bar chart, with each year's bar subdivided into portions by first contact date.
Many thanks AngleWyrm yes I understand we need to look at other factors to model the above I will keep your suggestions in mind. I don't have the data to map but what I am asking is what theory do I quote when I dispute that we should not look at the base on whole in such an instance as older customers have a establish relationship and some sorta loyalty, so they know the process and unlikely to make a complain etc to few more other details.