Text Data Mining
What is Text Data Mining?
Text data mining is the process of deriving high-quality
insights from text - often from huge volumes of material.
The process, using advanced 'machine learning' technology enables
the identification of patterns and trends that would otherwise
remain 'hidden' - because manual processes cannot accurately and/or
cost effectively extract the insights.
Ultimately, understanding and taking action on the insights
contained in customer comments can have a direct impact on
bottom-line performance, through identification of best practice,
integration with learning/development and more targeted
performance management.
What are the benefits?
Most forms of research include the gathering of comments, but
typically those comments are not analysed in a way that enables the
accurate identification of the insights contained
within them. For example, if promoters and detractors are
identified as part of the research methodology, what are the common
themes (and relative importance) that pertain to each group?
Depending on the way that text is analysed the benefits can
be enormous, including:
- What promoters have in common (the 'drivers' of positive
customer experiences)
- What detractors have in common (what needs to be avoided to
minimise dissatisfaction)
- What behaviours underpin positive and negative sales outcomes
(the'sales drivers')
- Geographic/demographic differences
- What aspects of the product or servicecustomers are responding
to in the most positive/negative way.
- What is building the value of your brand, and what is
detracting from it
Why should you invest in text data mining?
If you are already gathering customer comments, then you are
paying for the collection of that data. Without robust analysis
much of the value in collecting comments it is wasted. Text data
mining offers the opportunity to take huge volumes of your customer
comments for you to understand the trends/themes that would
otherwise remain hidden. The investment is controlled by the
frequency with which the analysis is undertaken, but even annual
analysis offers significant value. Where volumes are large, we
would recommend 2- 4 rounds of analysis on an annual basis.
