The power of graph technology for insurance price comparison

By Emil Eifrem | 8 March 2017

Emil Eifrem of Neo Technology explains how Wobi, Israel’s best known price comparison website for pensions and insurance, is using graph technology to enhance its service and stay ahead of the competition

Graph databases are helping the financial sector innovate new and exciting services because, unlike traditional relational databases, they are designed to join the dots and reveal the connections and relationships between data on a deep-seated level.

Of course graph databases are not a suitable solution for all the financial sector’s computing challenges. There are still transactional and analytical processing needs within financial institutions where relational database technology is well suited, such as systems of record in human resources, for instance.

But when it comes to price comparison websites, graph technology comes into its own. Financial price comparison websites are extremely popular with consumers as they let them compare a number of finance products, such as car, home and content insurance on one website, to make an informed decision before they buy without having to travel the Internet.

The world of financial price comparison websites is highly competitive and they are only as good as the results they give. Consumers will quickly move on if the service is not as accurate and efficient as they have come to expect. Price comparison websites therefore need to be able to generate compelling ‘value offers’ on an ongoing basis, and this can only be achieved if large amounts of consumer data can be analysed precisely and at speed.

Take the example of Israeli-based insurance and pensions leader Wobi. It is one of the best known price comparison websites of its kind in its market. It has adopted the power of graph technology to do exactly this and more.

It has 500,000 customers and millions of monthly visitors to its website and is already looking to expand into the banking and finance comparison service this coming year.

Wobi knew that to establish itself and entice customers back repeatedly it had to provide ‘value offers’. To do this efficiently it needed to create a detailed financial profile of each customer, including their family status, insurance policies, pensions, savings and accounts and ensure it was highly searchable, linking the necessary relationships to hook them into the most viable products.

The big data hurdle

After extensive research, Wobi opted for a single database solution capable of providing it with a granular-level picture of its customers. It was paramount that Wobi’s team could utilise a data engine that could take a deep dive into data on a customer when required.

However, the structure and type of data files that Wobi has stored had proved complex to work with. They are stored in a tree-like structure, which means that under each customer sits a mine of information. These large files made it difficult to get to information quickly and easily.

Wobi considered overcoming this challenge with an object database approach, but that did not work. Wobi started to look at other database approaches on the market and hit on graph technology, in the shape of Neo Technology’s Neo4j, as the best option for its business as it could work within the confines of its data fields, whilst delivering the searches it needed.

30 million relationships on show

Wobi’s business is about understanding the customer. The best way to do this is to build connections between them and their information, which is exactly what graph technology allows the company to do.

“Describing a relationship is important,” says Wobi’s CTO, Shai Bentin.“For example, we get new information about pensions monthly. To model this in an object-oriented way would be lengthy, whereas with a graph database I have a relationship called ‘latest’, and a relationship called ‘previous’, and I can just do a ‘select’ and get whatever I need.”

Wobi chose to run a graph database. It sits at the epicentre of Wobi’s network of 20 servers. Wobi now has a team of five people dedicated to graph database development and testing work to ensure reliable data is available at all times.

But where graph database technology is proving of real value is in tracking relationships that make data return value to a comparison website. The database currently runs around a million relationships a second and is handling half a million customers with an average of eight pensions, insurance policies and products. This equals a staggering total of 30 million relationships.

As a result of Wobi’s microscopic view on data, it can actually analyse customer data and advise them on ways they can save money on policies, utilities etc. “We can look at the customer’s account in such depth that we can tell them they must have a water leak somewhere because they have been paying way over the average for their water every month! We can also now easily see how to suggest a customer move from, say, one phone company to another, as the other supplier would better suits their actual expenditure,” he says.

The customer is king

Wobi’s story really does underline the power of graph technology in the world of price comparison websites. It is graph technology’s ability to read data that helps Wobi understand each individual customer. This is key to its success.

Companies like Wobi, who are ahead of their game, have turned to graph technology as the ultimate high performance database solution in the financial sector. Others will quickly follow their lead.

The author is co-founder and CEO of Neo Technology, the company behind Neo4j, the world’s leading graph database