How Suncorp is using data to prevent churn, predict claims

Suncorp Group has further bolstered its business intelligence army with 300 extra staff to help use information from its ever-growing data warehouse to stop customers from churning.

The group today committed an extra $75 million to the next stage of its transformation efforts, which it expects will reap $170 million worth of efficiency benefits by 2018.

The new ‘optimisation’ program contains four pillars, two of which centre on the group’s ‘Ignite’ core banking overhaul and its business intelligence efforts.

At its investor day this time last year, Suncorp revealed it had pulled together data analysts from its insurance and banking units to build an 850-person business intelligence centre.

It followed several years of work consolidating 12 data warehouses into a single IBM Netezza warehouse - what the bank is calling a “data lake” - to create a single source of truth on customers and build a “sense and respond capability”.

At its 2015 investor day today, Suncorp Group's business intelligence lead Mark Reinke detailed how far along in its journey the company was.

Over the past 12 months, the business intelligence unit was bolstered by a further 300 staff trained in modelling and analytics formerly scattered throughout the business, now working side by side in a dedicated function.

Suncorp also spent the past year adding more and more data into the lake - to a point where it now has 90 percent of its core data in the warehouse.

“And we’ll progressively add more, both external and internal data. We haven’t just put customer data in the lake, we’ve put all data,” Reinke said.

“We’ve put people, finance data - even last week we found an opportunity I didn’t even imagine we’d have three or four weeks ago: when you’ve got people data and bank transaction data in one place, we can look in 15 minute increments what transactions look like in every one of our branches and then match people and resources and workforce planning to that.”

Adding more core and external data to the lake will allow Suncorp to “hear customer signals in a way we can’t today”.

“It will mean the difference between traditional cross-selling and what we want to do: sense and respond,” Reinke said.

“We’re moving from a backward-looking capability to forward-looking. That’s not what’s happened in the past - we will be able to proactively manage what will happen.”

Suncorp also formed a partnership with data analytics company Mu Sigma to improve its ability to do more sophisticated analysis.

“We’re working on a portfolio of quick and sophisticated pilots involving machine learning, algorithmic learning and over time, spaces including artificial intelligence and a range of other things that we don’t have the capability to do at the moment,” he said.

One pilot involved using data to predict when a customer might leave, while another aimed to predict when a customer might need to make a claim for insurance.

In the first - dubbed “next best action” - the Suncorp BI team dipped into the data lake and used predictive modelling to test personalised interactions with AAMI personal insurance customers.

This approach - aimed at sensing what a customer might need - resulted in a 50 percent increase in policy cross-sell, and a 7.5 increase in renewal rates for high-value AAMI customers identified at high-risk of churn.

In the second pilot, Suncorp used data from the Bureau of Meteorology - contained in its data lake - during Cyclone Marcia earlier this year to track the cyclone’s predicted path.

It then determined policy holders who were likely to be impacted based on their location, as well as the likely extent of the damage.

It allowed the group to forecast the total sums insured and the likely cost of the claims - information which would normally take days to produce and verify, the group said.

The business intelligence team is also looking to further improve its “locational intelligence” so it can warn customers via SMS during a major weather event.

Outside of weather, it plans to also use predictive intelligence within its fraud, supply chain and claims operations.

Read on to learn about the results Suncorp is reaping in the lead up to its core banking go-live.

Early value derived from core banking overhaul

Suncorp Bank is “smack on schedule” to meet its promised July 1 2016 deadline to deliver its new Oracle-based core banking platform under Project Ignite.

The bank’s chief executive officer John Nesbitt first made the commitment this time last year.

The project involves rationalising the bank’s product set by half, decommissioning 12 legacy systems and re-engineering 580 processes.

The business decided to take a staggered approach to releases up until the go-live date, opting to tackle the least complex areas of the business first.

Personal loan origination was first in line, and Nesbitt today revealed the phase had been completed and the bank had processed more than 3500 personal loan applications since switching the system onto Ignite last November.

The entire branch network was now fully trained to operate on the platform, he said.

“This was a critical first step for the platform, as this low-risk part of our business allowed us to test over two-thirds of the systems across the bank,” Nesbitt told investors today.

“The staggered deployment of this release ensured a controlled rollout, and supported system functionality. It was a low-risk approach which provided lots of learning and validation across the system.”

The new core banking system is allowing Suncorp Bank staff to - in some cases - process personal customer loans in under 20 minutes.

“We’re now settling loans same day. This is a significant change from the pre-Ignite world,” Nesbitt said.

“Previously there were multiple touchpoints across our business in the approval and fulfilment of a personal loan. It took three to five days due to multiple hand offs. Not a very good customer experience.”

Suncorp Bank previously identified home loans, deposits and payments and business loans as next cabs off the rank once personal loan origination was complete.

Home loan origination will be shifted onto the new platform in November, Nesbitt said today, while term deposits and transaction account functionality will become available from the new system in early calender 2016.

“Once the upgrade to the new version 2.4 of the [Oracle] software is completed early next year, legacy systems conversion will start taking place,” Nesbitt said.

“We expect Oracle to deliver the simplified business banking system that we need by late 2015 so we can complete implementation of business banking by June next year.

“We’re very pleased with the progress of this project to date, but I can tell you we’re not complacent by any means.

"By their very nature, technology and change programs require absolutely diligence and vigilance, and we’ve got the right people with the right expertise both internally and externally supporting this program.”