The software, being trialled by London and Country, has been described by as a ‘catalyst for change’ in the protection market.
After the Mortgage Market Review elongated the process of applying for finance to buy a home or remortgage by several hours, selling protection, another cumbersome transaction, became increasingly difficult. As the mortgage market found its feet after the turbulence of the credit crisis began to settle, an adviser’s time became even more scarce – pushing protection sales even further to the margins of the mortgage sale.
Debbie Kennedy, group head of protection proposition strategy, said: “We decided to look at technology to solve the problem of declining protection sales. We wanted to make it easier to integrate protection into the mortgage process.”
Decision in seconds
Kennedy said the company was using a new kind of data technique which allowed it to look at existing Royal London customers across its book to build a statistical model of its expected applicants. Data such as age, salary and occupation, is collected during the mortgage application process and transported to a ‘black box’ which underwrites the protection policy.
A decision is returned within seconds. Using algorithms and analytics, the company can predict those customers who are likely to pay a standard premium and offer that rate to the mortgage client.
“By using this method, we have cut down the number of questions in a standard protection policy from around 30 to three which centre on the applicant’s medical history,” said Kennedy. “For example, we need to know if the applicant has had an overnight stay in hospital or if they are they under any special investigation.” By asking a few additional questions, Royal London has removed the need for a GP reference.
‘Change is coming’
Lucy Brown, head of protection sales, London & Country, said: “We are working in an industry where so little has changed. To work with Royal London on this project, really feels like change is coming.
“I think other providers will be clamouring to understand how they have achieved this and I’m sure it will act as a catalyst for change within industry.”
Royal London is currently working on applying the same technology to its critical illness policy which Kennedy admitted was ‘more challenging’. “We are in the testing phase at the moment,” she said.
Once both policies have been fully tested, Royal London will look to roll them out to the market.
Under the bonnet
For the tech-curious, Kennedy (pictured) explains how it works.
To build our statistical model we chose Microsoft R. It allows us to process the amount of business data involved and as it had the necessary machine learning libraries.
This statistical model, which predicts whether the applicant can be offered the standard premium, sits on a different server within an Analytics Decision Engine, Royal London’s formal name for it. Informally it’s ‘the black box’.
It has an API link that goes into L&C’s CRM system. API stands for Application Program Interface and refers to a set of a programming rules which allow two software programs to securely talk to each other.
London & Country requested a new web-service that couples their back-office application directly with the same scoring engine and underwriting rules, to allow its advisers to obtain instant eligibility feedback for their customers.