Open-ended Text Analysis Tools
One of the many tools available for examining data in the ViewPoint™ platform is dedicated to analysing open-ended textual responses. We always encourage the use of open-ended questions to get the Voice of the Customer feedback through the use of the ‘why?’ question.
Our text analysis engine now makes it much quicker and easier to see what respondents are saying in their text responses, even when there are huge volumes of survey responses. With these feedback tools there is now the ability to find answers to all the key questions.
Please contact us and let's explore the possibilities.
Categorisation or Theming
Finding the topics of interest to the respondents is done through the categorisation of the different textual responses. By analysing the articles of an open-ended answer it is possible to assess what respondents are concerned about, pleased with, or simply want to tell you about. What proportion of the respondents made a comment on a given aspect can then be calculated and displayed graphically.
For example, a pie chart can display what proportion (in percentage terms) of your respondents talked about cleanliness compared to other topics when responding. This analysis gives further indications as to what areas need looking into in detail and in particular can highlight areas that are not even mentioned in the quantitative feedback.
N-Gram Analysis...in brief
Word frequency, or more accurately n-gram frequency, is used as a basic tool for assessing and reporting the open-ended text responses. Having removed the ‘stop words’ (Words such as ‘the’, ‘and’, ‘I’) the occurrence of individual words (uni-gram), word pairs (bi-grams) and triplets (tri-grams) etc., is analysed and a frequency table calculated to show the words that are used most.
This is just for a taste of what people are talking about and should never be used in isolation, but gives a helpful indication of main topics with the texts. The actual frequencies can then be displayed in word clouds for a graphical representation, particularly useful for the presentation of results in coordination with other measures.
In simple terms this algorithm looks to inform you of whether people are being positive or negative in their text response. Various techniques are used to assess the feeling behind the text; what people are liking and disliking. The details behind all the decisions our sentiment engine makes is also available to us.
As with all the information held within the ViewPoint CX platform, the results of the sentiment analysis can be presented in tables or graphically using, for example, a doughnut chart.
Our Future with Text Analytics
The text analysis engine within the ViewPoint platform is constantly evolving, and in the future will include the use of neural nets and domain specific analysis to enable the system to learn more effectively about how you use it and how your customers communicate with you. New ways of analysing and presenting this invaluable information will regularly be added to the system to enable us to help our customers know their customers better.