Data Driven Customer Experience Management from A to Z

by Theo

Data Driven Customer Experience Management from A to Z

by Theo

by Theo

In this last blog in the series we want to use a customer case to illustrate all the steps involved, to give you a general feel for the entire Customer Journey Approach.


The customer use case

The case we are dealing with in this blog concerns an Insurance Company who wants to free up customer service personnel to be available as much as possible to answer compex  customer questions, the so-called ‘ Value calls‘. It is presumed that more straights forward questions, customers can take care of themselves through other (digital)  self service channels .However if a customer actually contacts service rep’s on these subjects anyway, these calls are classified as ‘hygiene calls’. The primairy goal of the Customer Journey Mining Exercise was to distinguish  value calls ‘ from ‘ hygiene calls‘. This in order to be more aware of the need for effective ‘hygiene solutions in the other customer contact channels. The ultimate goal being a better CX at lowers costs.


 Measures  and  KPIs

Customer Journey (management) always starts with drafting KPIs which contribute to the vision and strategy of the company. In this case identifying  ‘value calls and hygiene calls‘ was the core objective. A key KPI is the FCR (First Contact Resolution) and FTR (First Time Right). The definitions are not the same everywhere, but this KPI ensures a customer question is addressed at the first attempt. The FTR applies within a particular channel (i.e. customer service) whereas the other KPI is omnichannel. For example, did the customer try it online first and then make a call? Eventually CS (Customer satisfaction) over the entire contact is also a key KPI used to  be able to measure the customer’s total CX perspective. Finally, it is also important to look at the content of the calls. This information is required anyway to calculate the FCR/FTR and you can also set threshholds (absolute or relative) . For example, a target of a maximum of  500 calls per month via the My Login . Or only 10% of all incoming calls about the My Login .


Customer Experience as the foundation of the organisation

This use case started with a small project team with a multidisciplinary character. The project lead came from the department: Business Change & Customer Satisfaction. Data (insights) were leading from the outset, which means that a large part of the team, from different perspectives, have affinity with data. Certainly in the beginning the ambition was to keep the scope and  number of people required to a minimum. This with a clear set of KPIs, made the approach  clear to all concerned. Once the pilot was finished the project was adopted by the organisation and the project team started with a subsequent sprint. 

Data Sources and channels as input for Journey Mining

Initially it is important to understand which customer service  information is actually recorded. Often there is a CRM system where the employees log phone calls. In some cases there is also a link to a knowledge management system in which the various discussion topics are categorised in a structured manner. Even if this link does not exist , the  knowledge management system can be an important source. Finally every customer service system involves a telephone system. This can help to understand the handling times of the various subjects, especially if a link is possible to the aforementioned systems. The conversations can then be converted into a transcript , which may complement or replace the logging entry  in the CRM system.


Customer Journey Mining

Knowing what the context around a question is can be of great help in resolving the nature of the question. What was the customer doing? What was the  intention? Who did they want to reach out to? The answer lies in constructing customer journeys based on data. To do this you can combine data already collected,  with data from other channels (e.g. email and chat) , but also data from the back-office (such as transactions and requests).

Next you want to establish the relationship between all data points collected. This means classifying events and grouping activities and episodes to get a clear picture of the entire customer journey.


The role of Text Mining in Customer Journey Mining

What greatly helps in classifying events, is collecting as much meta-data as  possible from these events. So using not only information that has been logged but also information that you can deduce from other data sources. Many of the meta-data is hidden in the text fields provided by employees, but even more importantly, by customers themselves in feedback forms. Interesting aspects you can search, for example, are life events or stated  emotions.

The methodology used to transform free texts into  structured texts is called Text Mining Analysis. This is a complex world but what it boils down to is how you get from unstructured to structured texts. This involves searching certain combinations of words and the texts must first be properly pretreated . For a comprehensive overview of  this topic we  refer you to blog number 4.


Customer Journey Mining: several methods

To assist you in making proper interpretations of the context surrounding events you can use customer journeys or process descriptions. This is a  top-down approach and it helps to target the search for related events and relationships. The down side here is that you will not find unexpected customer journeys because you won’t know where to look for them.

The opposite approach is letting the data speak for itself by gathering all information around events to then discover the relevant events, is also an option. The difficulty here is that events could be attributed to multiple journeys, or it could be the case events have not been logged properly and information is missing (in for example the CRM system). The upside is that it is possible to discover atlernative routes which might not have been originally intended. Getting experts to validate this bottom-up approach adds much value , they can help to interprete the data and to confirm the validity of the results.


Whichever approach is followed, a lot of data is collected. To turn this data into clear actionable insights for analysists and other end users, it is recommended to visualise this data in intuitive and attractive visuals. In the case of the insurance company, you can chose a treemap in which the different subjects are visible in blocks, where the frequency determines the size and the colour the ratio between ‘value’ and ‘hygiene’calls.

Another form of visualisation  is a heatmap where the subjects are plotted against the customer Journey, i.e. the underlying episodes and within that,  the underlying activities .

The way in which the information is going to be used in the business will determine which type of visualisation is optimal.

After the succesfull implementation of a project, it is time for handover to the business. Customer Experience Management is a continuous iterative process and the main prerequisite for success is that the results are embraced by the colleagues  who have to work with customer journeys on a daily basis!

Leave a Reply

Your email address will not be published. Required fields are marked *