Just because you find correlation in your data when analyzing survey results does not mean there is causation. I find this is a common mistake in lots of survey reports when someone is new to survey research or conducting data analysis. Take this example:
When conducting the survey analysis of the survey responses, you find almost all clients who are contacted every few weeks are very satisfied, but clients who are rarely or never contacted are very dissatisfied.
Some people see this connection as a causation. Customers are satisfied because you contact them frequently to make sure everything is going well. The problem is, it's not a causation. Causation are extremely hard to prove because you cannot control every factor. For example, you may split your territory by industry and your solution suits some industries better than others. Or clients who are really satisfied simply use the product more often so the customer success managers reach out to them more frequently, because they are more likely to have questions. While those who use the product less have less to be satisfied about and may feel they are wasting those budget dollars.
To be able to prove causation, you need to be able to rule out all other possible explanations for the connection. As you can imagine, that's almost impossible to do since we do not control outside factors influencing the survey respondent or even the greater survey sample. Instead, when situations like these occur, we're seeing a correlation between two things. In my customer survey question examples, there's a correlation between how satisfied customers and how often they are contacted.
This principle does not only apply to customer survey research, it also applies to analyzing employee feedback forms, product surveys, market research and any other type of data collection and analysis.
Your organization sells products and services in the business-to-business space. As part of your model, each organization has a customer success manager who is responsible for reaching out to clients and ensuring they're using the product appropriately and ensure they're satisfied customers so they continue to be customers. As a result, it's really important to continually measure customer satisfaction. To do this, you've purchased customer feedback software to conduct customer surveys.
You follow all the survey best practices and keep your survey short. Two survey questions that are always asked, for example, are:
You follow all the survey best practices and keep your survey short. Two survey questions that are always asked, for example, are:
How satisfied are you with our products?
How often does your customer success manager reach out to you?
How often does your customer success manager reach out to you?
When conducting the survey analysis of the survey responses, you find almost all clients who are contacted every few weeks are very satisfied, but clients who are rarely or never contacted are very dissatisfied.
Some people see this connection as a causation. Customers are satisfied because you contact them frequently to make sure everything is going well. The problem is, it's not a causation. Causation are extremely hard to prove because you cannot control every factor. For example, you may split your territory by industry and your solution suits some industries better than others. Or clients who are really satisfied simply use the product more often so the customer success managers reach out to them more frequently, because they are more likely to have questions. While those who use the product less have less to be satisfied about and may feel they are wasting those budget dollars.
To be able to prove causation, you need to be able to rule out all other possible explanations for the connection. As you can imagine, that's almost impossible to do since we do not control outside factors influencing the survey respondent or even the greater survey sample. Instead, when situations like these occur, we're seeing a correlation between two things. In my customer survey question examples, there's a correlation between how satisfied customers and how often they are contacted.
This principle does not only apply to customer survey research, it also applies to analyzing employee feedback forms, product surveys, market research and any other type of data collection and analysis.
Powered by Compendium Blogware
Comments for Difference Between Causation vs. Correlation in Survey Data