Commercial organisations are constantly incorporating external data sources to enrich their view of clients, opportunities, and competitors. The quality of data procured determines market intelligence, and therefore the decisions a company makes. Those who have the best intelligence are set to be in a position of strength.
Managing the consistency and integrity of information can become an arduous task. If internal data issues exist, they may be exacerbated when trying to scale. 21st century business models don’t work well with siloed legacy platforms and poor-quality information. What would happen if your organisation decided to complement its existing data with external data that may be several degrees greater in volume?
The 2021 New Vantage Partners Big Data and AI Executive Survey questioned the focus on data within 85 firms classed as ‘industry leading’ or part of the Fortune 1000. Responses show that:
- Only 48.5% are driving innovation with data
- Only 41.2% are competing on analytics
- Only 39.3% are managing data as a business asset
- Only 30.0% have a well-articulated data strategy for their company
- Only 29.2% are experiencing transformation business outcomes
- Only 24.4% have forged a data culture
- Only 24.0% have created a data-driven organisation.
The issues around data and operations, if resolved, can deliver substantial ROI. Resolution is often found through an effective strategy that prioritises a high standard of data assets whilst complimenting existing business objectives.
Digital business models lose their advantage without trusted, reliable and relevant data. How do businesses ensure their current strategy puts them in a position that’s ready to scale and integrate third-party knowledge?
As the insurance sector has increasingly digitised operations, it has also seen a growth in data roles created and filled. The emergence of a Chief Data Officer (CDO) emphasises the importance of an effective data strategy on organisational goals. Data strategy in this case is defined as: “A coherent strategy for organizing, governing, analyzing, and deploying an organization’s information assets that can be applied across industries and levels of data maturity.” - DalleMulle and Davenport, Harvard Business Review
Data Governance is key to an effective data strategy. It provides a framework to adequately manage legal and compliance obligations whilst focussing on business growth. This is vital when the appetite for data is growing due to ambitious goals and market pressures.
“To me, legal and compliance are an absolute requirement but not sufficient. Why collect and store data if it isn’t going to drive business growth? This makes data a liability rather than the asset that it should be. For an organisation to unlock the value of the data it holds, the ethos across the firm needs to shift into ‘how do we maximise the use of all of our available data to increase revenue and reduce cost and risk?’.” - David Riley, Head of Business Development at Alqami
A core discipline of Data Governance is Data Management, which strives to curate accurate information with high degrees of integrity. Integrity in this case means we can confidently trust and interpret results from the data as ‘true facts’ (or a golden source). For this to be the case, we would have to know the data’s origin, that it’s up to date, and it corroborates with other internal/external sources.
An effective data strategy backed by both data governance and management should yield a clean and accurate set of assets that are ready to be enriched by third party sources.
Fig 1: Internal Data Standardisation
The utilisation of external data is a way to unleash the latent value and potential of your organisations existing knowledge. Internal data reflects an organisations own experience, whilst external data reflects the experience of the third party. Combining data pools gives a richer understanding of the market, adding perspective to an existing position. This is typically achieved through multiple datasets co-existing in a common data model. It’s like reading a book, where you gain a lifetime of experience in a matter of hours. However, not all data is created equal, and it’s important not to sacrifice the standards achieved from an effective internal strategy.
Last year, the data strategy team at Concirrus began the ambitious journey of quantifying cargo shipment exposure at an international scale. To do this, we had to know everything relating to good type, origin, transport method, sender, and recipient. There’s no single acquirable, standardised source of this data. It’s not surprising that given $19.48tn of international merchandise trade (WTO) in 2018 alone, it’s quite a hard thing to measure.
However, we did find that rich pockets of information exist in isolation, originating from customs sources or from the logistics industry itself. Yet, there were still challenges:
- The data is collected from manually keyed sources
- The data can be self-conflicting
- Each source of data is idiosyncratic in what it reports and respective format
How do we resolve these issues? Let’s look at two distinct sources of Chilean shipment data as an example. Both sources show an export shipment of fine Malbec on the same date. With a human eye, we can discern that:
- The companies in the transaction look similar by name.
- The goods descriptions are also similar. One is in Spanish which can be translated, and the other has been translated already from the source.
- The port of lading is slightly misspelt in one, but not enough to omit.
- The name of the vessel that carried the goods differed in each case but resolvable through cross referencing our existing marine data to prove which vessel was in port at the time.
- The values of the goods are similar when converted from source currency to USD.
This audit tells us we can be confident that both sources are referencing the same transaction. However, there’s over a 1.6 billion records to process (growing at a rate of 22.4 million each month), which is highly inefficient to do manually. Combining both Data Governance and Data Science disciplines allowed us to master such records at such scale. We did this by:
Defining a taxonomy and information model which
- Unified un-standardised sources of supply chain information into cohesive entities
- and provided the ability to join, relate and model internal insurance information in a common manner
Establishing a set of tools and methods which at scale efficiently
- Clean, transform, standardise, cross-validate, and enrich external supply chain data with other third-party sources.
- and create a set of ‘master records’ combined with our internal data which we can trust as ‘true fact’
This sequence can be applied to fragmented datasets to draw a collective pool of third-party intelligence. Doing so derives data that matches the state of internal datasets which are subject to an effective data strategy.
Fig 2: External Data Standardisation
It’s here where InsurTech’s such as Concirrus add real value outside, continually amassing and curating external data from a multitude of sources into a single pool that’s ready to bind with internal insurance data assets. The integrity of this pool determines the value of insight drawn from further modelling. Concirrus’ dataset spans over 2 Trillion data points and is continually growing, amassing new datasets from verified partners around the world. The efficiency savings from accessing the data pool alone are extensive. When combined with Concirrus’ modelling capabilities, such as pricing, the value-add scales further.
The single source of truth
The influence of Data Governance and Data Management on Data Strategy is what ensures datasets are ready for modelling. Applied to both internal and external data sources, these processes set a standard by working together to ensure data from multiple environments can be unified into a single source of truth that can continually scale. The maintenance of such processes ensures that regardless of origin, new data streams can be integrated and contribute to the growing intelligence of an organisation. If the integrity of this source is sustained, models will draw actionable insight for fast and effective business decisions that compliment modern business strategies.
Fig 3: The creation of a single source of truth