Text Mining is still in fashion

Two technological companies, regarded as some of the biggest ones in the world, ‘make their living’ from the extensive knowledge about their users. Google and Facebook obtain 80 and 95 percent of their respective revenue from ads. Such results are achieved thanks to the data that allows them to determine which ads have a positive impact on  individual users.

It’s not only the Silicon Valley giants who can use the data of their clients for revenue generation. According to McKinsey, a consulting firm, results achieved by companies that analyze client behaviour are 85% better and their gross profit margins are 25% higher.

Make Omnichannel work for you

Google and Facebook’s advertising businesses are successful because these companies collect and analyse even the smallest pieces of information concerning their users. Omnichannel communication provides similar possibilities but on a slightly smaller scale. A short historical analysis will make it possible to understand it.

Origins of retail trade: a shop owner relies on questionnaires and feedback from employees to obtain an image of who his clients are.

Origins of delivery sales: telephone customer service enables to create a client’s image with collected shopping history, preferences, and complaints.

Origins of loyalty programs: loyalty programs fuelled by data enable client segmentation.

Nowadays, every company has megabytes of data about every Client at their disposal that come from:

● phone call recordings,
● e-mail correspondence,
● social media interactions,
● tracking website clicks, shopping history, receipts etc.

The so-called Text Mining is ideally suited to analyse the data mentioned above.

Introduction to ‘Text Mining’

Computers require data in predictable formats. This is what distinguishes us from machines. For example, we can understand a sentence which is not grammatically correct. ‘Text Mining’ allows computers to separate useful data from unordered human language. It is a form of analysis using linguistics, statistical methods and machine learning to transform the text into data which is ordered and computer-friendly.

Thanks to ‘Text Mining’, the computer can download a text and analyse it in terms of its content, sentiments, categorization, summaries and other elements. Let’s focus on the fact of how few companies apply such an analysis of their communication with clients. Omnichannel communication strategy means channelling the whole interaction with the client into one place. What used to be dispersed across many systems in the multi-channel world is now easy to analyse.

What can ‘Text Mining’ do for you?

At your disposal you have many e-mails concerning your communication with the client, automatically generated transcripts of phone call recordings, text messages, internet chats, interactions with Internet bots, shopping history etc. – what can you do with that?

Text mining solutions will enable the analysis of historical interactions, will affect the improvement of subsequent interactions with the clients and will enable automation. Here are a few examples:

History analysis

Detecting the so-called failure points within the ‘customer journey’. Do clients often get in touch with you in specific situations and why? The customers’ bad experiences. Text analysis may help with determining the reasons behind the client’s dissatisfaction, which allows the introduction of necessary future changes. Agent’s/Seller’s efficiency assessment. Instead of listening to one of a hundred recordings, it is possible to analyse all of them immediately, draw conclusions and prepare appropriate training sessions for the agents etc.

Future improvement

Forecasting client behaviour. ‘Text Mining’ will make it possible to understand whether certain types of interactions, manners of expression or sentence structuring lead to a specified result such as making a purchase or cancelling service.

Load forecasting. Analysis of the whole range of communication with the clients will make it possible to scale an appropriate customer support centre on a long-term basis.


Request managment. Thanks to the semantic analysis and categorization, a particular text may automatically trigger certain actions on the part of the customer support centre. This possibility shall be viewed as some form of IVR for Omnichannel.

Internet bots. ‘Text Mining’ is an important part of the process which should enable bots to ‘understand’ and react accordingly to the human language.

Thanks to the ‘Text Mining’ technology, the execution of the above examples may be automated, which reduces the cost and enables real-time access to data and analyses.

Law and ethics

In the era of GDPR and general consumer concerns related to personal data, it is important to find the right balance between respect for one’s rights and the development of a more effective business.

Luckily, most applications of ‘Text Mining’ make sense even when we’re working with anonymous data. For instance, searching sentiments in client communication history is valuable even when we don’t know their names and surnames.

If the personal data constitutes a crucial part of the data or you can provide better customer support through the analysis of individual data, you need to make sure that your privacy policy allows such actions.

The situation becomes more complicated when you use data that is publicly available and you are trying to associate it with the details you have about your clients. And how about the situation when you’re analysing the social media data to obtain additional information about their lifestyle? For example, when a tool such as Text Mining can see that the client often posts comments about Legia Warsaw football team, should the call centre agent mention the last game they won in his next conversation with the Client? 🙂

Even when something is legal, it shall be considered whether the clients will appreciate this or this would have the opposite effect.