Moving into a year of financial uncertainty, placing the customer as the focus of your information landscape is critical for targeted growth. By creating a solution where the customer is the primary measurable metric, across all other dimensions, we create
a view about the customer, from every aspect of the business and unleash insights faster. From premiums paid, claims made and products bought, we can analyse client habits, and optimise the journey for both the customer and the organisation.
That journey needs to be optimised for each unique customer, and with the increase in vulnerable customers, brought on by COVID-19, having a means to quickly ingest and analyse data for insights, such as customer characteristics, is critical. In creating
an end-to-end customer 360° view data model, we are able to do just that. The data model focuses on customer datasets, which are able to be directly used for predictive insights, such as customer churn, segmentation, hyper personalisation, or even, the identification
of vulnerable customers.
The 360° view
The idea of this view is that by understanding the customer across a multitude of dimensions, we not only generate insights from their behaviour and spending habits, but can then enhance the customer experience, so they continue to want to engage further
with the organisation. So, how is this different to traditional data modelling where the customer is one ‘dimension’ associated with items such as premium transactions (which also have numerous other dimensions)?
The difference here is that the focus is now directly on the customer themselves, and the datasets that are produced are useable inputs into analytical models. Customer detail is the key identifier and unique information across all of the datasets produced.
Those customer identifiable datasets are broken down into the data that is available within an organisation, and other data which if it could be obtained would be very useful in order to create a more holistic view of the customer.
So, breaking down customer 360° into datasets of potential insight, we need to consider numerous options, which can include:
- reference views (non-analytical)
- customer master details (primary reference for subsequent processing)
- reference views (analytical)
- customer policy and product
- customer claims
- customer premiums
- customer finance transactions
- customer personal finance
- customer sensory / IoT (wearable health trackers, or smart cars)
- customer and company social media
- customer company communication interaction and channels
- output customer classification and insights
Customer datasets will continue to grow
This is not an extensive, nor an achievably possible, list. The number of customer datasets will continue to grow as new data becomes available. For example, if an insurer is not using smart devices to monitor customer health and behaviour today, IoT datasets
will not currently be available, but they may well be in the future. Similarly, if an insurer does not have access to customer banking spend or financial habits today, they may well do so in the future through mainstream adoption of open banking.
The customer 360° view is not a replacement for the existing data transformation programme, unless the transformation is to re-design the estate in full. Therefore, although a customer 360° view may be an architectural design or platform within the organisation
(which the enterprise might like to take advantage of), certain business units – such as the finance team – will always have a focus on creating financial statements, rather than a customer view of the data. Creating this view can still be done independently,
or as an enhancement, to the existing enterprise (which could be a traditional data modelling architecture), regardless of the current situation.
What is the difference between traditional data modelling techniques and a customer 360° view? To understand this, we must understand that there are a multitude of ways to design data models. The traditional methods focus on either normalising or de-normalising
data, into relational or non-relational objects of data.