Sources of Data for Building a Customer360
In this blog we’ll look at the different sources of data available to build a Customer360 solution.
The move from being Policy centric to Customer centric is being undertaken by most Insurance companies and often forms part of Omni-Channel, Customer Centricity or Digital Transformation programmes.
The goal of constructing and operating a Customer360 isn’t an easy one to achieve given the challenges around finding, accessing, integrating, matching and merging of data from mostly Policy or Product oriented systems. What makes this approach even more challenging is that there is a great deal of other sources of Customer related data that is not part of a traditional Policy oriented environment. This data helps create a much wider, much richer and more relevant view of a Customer that enables Insurers to create more targeted marketing offers and generate cross-sell/up-sell opportunities as well as improve service delivery.
In this blog I look at some ideas on the potential sources of data used to construct a Customer360 in Insurance, initially in broad terms but with some specific ideas around types of data.
One question I get asked frequently is about how to start constructing a Customer360 and what it will look like at the end. I tell Insurers one way to think about building a Customer360 is to think about it in 4 phases. Each phase equates to 90 degrees on a compass and each subsequent phase builds upon the previous one. So we start by building a Customer90, build more creating a Customer180, build more still creating a Customer270 and finally the Customer360.
This isn’t meant to be a project plan nor is it the only way of doing this. It’s more a way of thinking about the possible types of data being utilised as part of a Customer360 solution and how together they contribute to the wider, richer picture of a Customer that we’re looking for.
Phase 1: Building a Customer90 with internal, structured data
For most Insurers this is the obvious starting point. In this phase we’re looking to start the construction by utilising structured data that sits within the confines of the organisation. Structured data is data that is typically stored in a form where there is known definition of what the data is and how to access it. Typically this type of data is stored in databases, applications and files.
Most Policy systems capture significant amounts of structured data such as name, address, contact details, product type, payments etc. Underwriting and Claims systems also hold structured information such as risk ratings or payment details. Contact Centre systems capture and store information about date and time a Customer called, problem or request type, resolution codes etc.
All this structured information is valuable in that when it is brought together as part of the Customer 90 as it makes core information about the Customer, and their basic relationship with the Insurer, available.
Creating the Customer90 view will provide an Insurer with enough information to understand their basic relationship with a Customer and then master that data on an on-going basis. It will tell them things like what products a Customer bought, if there are any outstanding claims or if there are any outstanding issues. This helps the Customer Services and Marketing teams getting a better understanding of the relationship with a Customer so they can interact and market more appropriately.
Phase 2: Building a Customer180 by adding external structured data
There are a range of external sources of structured information that can add useful insights into Customer data.
There are a range of third party data services vendors, who can provide data around credit scoring, professional qualifications and affiliations, broader financial profiling and various group memberships.
Additional insights around businesses can be obtained from various providers of business demographic and business makeup which can be blended with existing internal Customer data. This will provide a broader view of a business Customer to support more targeted segmentation.
If an Insurer is in partnership with third party vendors around Internet of Things (IoT) devices, such as vehicle black boxes or fitness trackers, then the structured data generated by these devices can also be blended with the core Customer record.
And then there are a number of commercial and governmental Open Data sources that can provide useful market and demographic data, even though it’s anonymised, which can also add value.
Phase 3: Building a Customer270 by adding internal, unstructured content
Most Insurers have a wide range of unstructured data sources that can expose a wide range of Customer insights when processed accordingly. Unstructured content is data that has significant aspects to it that have no set definition. An example would be an image file – we know the file type and how to process that but have no way of knowing what the image is of as this could be virtually anything.
The first challenge is to identify the source of the unstructured data and develop a method of processing it to unlock its insights. This is an area when many of the Big Data processing technologies, such as Hadoop, are used to create structured insights which can be added to a Customer record.
Examples could include:
- Apply text analytics and language processing technologies to electronic forms from Customers to extract insights from within written text. The electronic form may contain fields of data that allow Customers to freely write text so the text processing can mine this text for additional content such as claim details, change of circumstances details, authorisation & delegation requests or complaint details
- Scanning of structured paper forms from Customers into an electronic format and applying optical and intelligent character recognition can create electronic content which can be mined using text processing technologies
- Application logs, such as those generated by a Customer facing application, capture a great deal of information about how the Customer has physically interacted with the online systems of an Insurer. These log files can be parsed using log file parsing tools (i.e. RegEx) to generate granular information about Customer activity.
- For example they can look at the logs for a mobile app and work out what parts of the app the Customer interacted with and how they interacted with it. This can generate preference information that will help understand better ways of communicating with Customers
- Speech-to-text technologies can take recordings of Customer conversations and generate electronic documents, which in turn are parsed using text analytics tools to generate insight.
Insurers typically hold significant amounts of unstructured content which, when properly processed, can yield previously unknown insights about Customer behaviour.
Phase 4: Building a Customer360 by adding external, unstructured content
The final phase is probably the most complex as it usually involves either content that is very unstructured (such as written Social Media posts) or content which is inherently hard to process (such as moving video).
This situation can become more complex when the unstructured content may contain information that requires significant additional processing. Examples of this could include social media commentary using shorthand and abbreviations or text that has domain specific content that has content which is meaningless outside that industry (i.e. medical notes).
Regardless of the complexity there are useful insights to be acquired if the content can be parsed. Social Media posts can detect sentiment which can show how a Customer feels about a product or service without the Insurer having to physically get in content with them. Images can show relative vehicle positions to support claims processing activities after a motor accident. Video imagery can show the sequence of a motor vehicle incident as captured by an in-car recording system.
Whilst potentially complex to process, these sources can yield insights that would be unobtainable otherwise.
Enabling Capabilities to build the Customer360
At each Phase there is a need for some enabling capabilities that take data and relate it to a specific Customer. This type of enabling capability is called Master Data Management (MDM) and is a set of technologies that uses sophisticated data processing techniques to examine data and decide whether a data record is related to a specific Customer or not. This type of data processing technique is useful in looking at large amounts of data and deciding whether data records are associated with a specific Customer and if so, creates a link between the two. For many Insurers, especially those undertaking Omni-Channel transformation, there is also a need to accomplish all this in near real-time as in this environment a Customer can switch channel very rapidly and they still expect the Insurer to know everything about their experience even if it was initially through a different channel.
This processing enables the MDM solution to create a high quality, trusted view of each Customer that is built up from a wide range of internal and external data sources. It incorporates insights from the phases outlined above to become the single, authoritative source for Customer data which is wide in scope and rich in depth.
Each phase contributes to the width and depth of the single, authoritative source for Customer data and, when combined together, provides huge potential for the generation of new insights that can be used across the enterprise.