Die Bayerische is using the world’s leading graph database, Neo4j, to provide consistent information on customers and contracts. Neo’s CEO Emil Eifrem discusses the issues the customer faced and how graphs have helped

Graph databases have gained a growing reputation in the financial services sector. Developed in-house by the consumer Web giants at the end of the 1990s: Google, for instance, using graphs, exploited the connections in every Web document to rank search results in order to corner its market. Now these technologies are available to the wider market.

Unlike most other ways of looking at data, graph databases are designed to exploit relationships in data, which means they can uncover patterns difficult to detect using traditional representations such as tables. As a result, Forrester^[1], says over a quarter of enterprises will be using such databases by 2017, for instance, while Gartner believes that over 70% of leading companies will be piloting a graph database by 2018[2].

There are a huge number of important applications that require the sort of high performance and focus on relationships graph databases provide. Consider onboarding a new trader and programming the Web of parameters.

Similarly, graph databases can increasingly use relationships structures to identify fraudulent transactions that might not meet classic fraud criteria.

One such example of a financial services company needing a high-performant database to manage a large amount of connected information is Munich-based insurance leader die Bayerische, one of the most specialised, medium-sized insurers in Germany. For a century and a half, its comprehensive range of insurance and financial solutions have been a favourite among savvy customers in its native Bavaria and beyond.

With a brand promise of ‘Insurance with the Reinheitsgrebot’ (the famous mark of quality for German beer), the aim has always been to maintain competitive advantage through superior services. And key to its latest project on this front is inventory database information available at all customer contact points.

For insurance companies like die Bayerische, a heavily heterogeneous system landscape for managing insurance contracts is simply a fact of life. There’s always going to be a multiplicity of back-end data systems that need to be maintained; after all, pension insurance or disability insurance can last for several decades. The drawback is the on-going effort needed to provide consistent information on customers and contracts in a consistent internal business user interface.

To manage this situation better, this major European financial services operator recently brought in local service provider Intelligent Solution Services AG (iS2) to build a new, Internet-based field staff system in order to improve availability and performance at the point of sale.

The background to that decision is significant as iS2 has a solid track record supplying a large proportion of German financial service providers sales teams with electronic price lists – and its main focus is consulting, design and implementation of sales-oriented software solutions for this critical vertical.

Single version of the truth

In die Bayerische’s case, it soon became clear that to meet its needs, customer data needed to be replicated from different host systems and regularly synchronised. If the team could manage this, it would result in a standardised data framework that would provide the customer’s knowledge workers with a single version of the truth – the Holy Grail of all BI (Business Intelligence).

The problem, however, was that traditional relational databases were not going to provide all the functionality required. Why? Because, says iS2, the sheer volume of data was too large, while running endless JOIN operations for queries was going to be too computationally expensive.

Alternative database systems were therefore required – and as the VAA (Voting Advice Application) model used in many insurance products is similar to a graph model, the use of a graph database was an obvious next step. That’s because, says the German IT firm’s MD, Thomas Wolf, ”A graph database can perfectly represent the products of an insurance company; contracts are linked to people and these people are linked to other people with other contracts. This leads to a compact data network that can be quickly and easily navigated or traversed. All related information can be retrieved from any starting point quickly too.”

Responding in a few seconds

After an intensive evaluation of its technical options here, iS2 decided that market-leading graph database platform Neo4j would give it the best basis for the die Bayerische project. The database especially impressed with its data model’s richness and flexibility and whiteboard friendliness that it says facilitates discussions between developers, as well as a shallow learning curve.

After a successful proof of concept, iS2 steadily developed the new customer system with the support of the firm behind Neo4j, Neo Technology. One of the challenges that had to be faced early on was adequately representing the quantity of products and formats in the customer’s extensive financial product database. The flexibility of the basic graph model quickly resolved this issue of volume, however.

The end result of the partnership between iS2, the customer and Neo is an application called Bay4all, which enables insurance advisors, call centre employees and insurance brokers to access all relevant customer and contract data with good performance levels. Now in full production, approximately 400 insurance employees use the new information system.

This new field staff solution is also supporting die Bayerische employees with speed, availability and consistency – when querying information, response times are on average between one and two seconds, which is delighting both customers and the staff helping them.

The insurance company’s inventory systems are also synchronising with Neo4j. Data buffering, for example, means die Bayerische teams can access the system at any time and it can be restarted at short notice in the event of an error.

Flexibility and scalability

Looking back on the project, there’s only one conclusion: ”Neo4j was the right solution for us,” Wolf confirms.

“It’s a fully developed and stable database which supports us very well, thanks to its query performance, flexibility and scalability – exactly what we needed to meet our clients’ requirements.”

The verdict’s clear. Experiences like die Bayerische’s and many others underline how large is the potential contribution of the graph approach for any high-performance database need, where information needs to be available at any time at all customer contact points and offering high performance, flexibility and scalability.

Die Bayerische and other large insurers and banks are benefiting from Graph and so should you.

The author is co-founder and CEO of Neo Technology, the company behind Neo4j, the world’s leading graph database (http://neo4j.com/)

[1] https://www.forrester.com/Market+Overview+Graph+Databases/fulltext/-/E-res121473

[2] https://www.gartner.com/doc/3100219/making-big-data-normal-graph