How Numr CXM used Text Analytics to find high impact attributes from feedback.
Our client is a leading insurance provider in the country. They were routinely collecting customer feedback through NPS® surveys. However, they were unable to perform feedback analysis of unstructured comments. This is problematic since the NPS® score is merely a number. Feedback comments are more crucial since they provide the reasons behind customers’ ratings.
Mainly, Numr faced 3 challenges. These were-
(Say, if a customer has given an 8 on the NPS® scale, our challenge was to explain the reason WHY)
2. To extract KEY DRIVERS/ AREAS that have the highest impact on customer satisfaction (recommendation)
3. To establish a relationship between the NPS® score and feedback comments.
We analysed around 3000 comments using our Comment Categorisation Formula to arrive at extremely precise conclusions. Our process was as follows.
STEP 1- CREATE SEVERAL LEVELS OF DRILL DOWNS TO CATEGORISE ALL COMMENTS
To run feedback analysis, we created several levels of drill downs.
LEVEL 1– was the overall, broad comment category. Basically, we categorised/ separated all comments into either POSITIVE or NEGATIVE.
LEVEL 2– represents sub-categories/ areas, such as, ‘Customer Care’, ‘Policy Features’ et cetera.
See the chart below to understand how we create several drill down levels for comment analysis.
STEP 2- IDENTIFY KEY AREAS/ DRIVERS THAT HAVE THE HIGHEST IMPACT ON RECOMMENDATION
We used Text Analytics to automatically classify all comments into appropriate sub-categories.
Then, we used Regression Analysis to isolate the impact (in terms of a β-value) an issue/sub-category will have on the overall NPS score.
For instance, see the POSITIVE COMMENTS chart below
We identified 9 main areas/sub-categories.
Now, each bubble represents one of the categories mentioned above.
The X-axis represents the % of Promoters (customers who gave a rating of 9 or 10 on the NPS scale)
The Y-axis represents the IMPACT the sub-category has on the NPS score
And, the size of the bubbles represents the volume of comments under that category.
We asked our client to prioritise areas with the highest X-axis and Y-axis value.
FINDING 1- ‘Customer Care’ and ‘Customer Experience’ have the highest positive impact on Customer Experience. (We have circled them on the Chart)
Similarly, for the NEGATIVE COMMENTS CATEGORY, we followed the same process and discovered that
FINDING 2- ‘Follow ups and Communication’ and ‘Documentation’ were making customers the angriest. Since they had the highest negative impact on Recommendation, we advised our client to prioritise them.
What’s more, we broke down larger categories (volume of comments) into sub-categories for a granular and precise analysis.
For instance, we further broke down ‘Time taken for Approval’ into ‘Initial Approval’ and ‘Final Approval’ to discover the EXACT source of the problem.
See the chart below for details.
Our analysis allowed our client to obtain granular insights that were infinitely actionable in a very short period of time.
With Numr’s help they were able to accurately pin-point specific issues and prioritise KEY AREAS that were having the maximum impact on customer satisfaction.
Which, in turn allowed our client to serve their customers better and finally, boost Profits.