How Suncorp is using analytics to improve customer outcomes
A combination of geospatial technology, AI and machine learning helped Suncorp win the first-ever Practice Prize for outstanding applications of business analytics.
But the team behind the innovation – pricing managers Emily Chong (pictured, centre left) and Steven Farrugia (far left), data scientist Jim Hoang (centre right) and executive manager Nagender Chetti – say they’re just getting started.
“It's fascinating to see that Suncorp is recognised for its leadership in data and ethics and how to bring the value from data analytics to the customer,” says Jim.
According to Emily, the prize has boosted the team’s morale, showcased their analytics capability with the wider community and helped to attract and retain talent.
Experimenting with Google Street view
During their presentation at the Melbourne Business Analytics Conference last October, the team shared how they analysed satellite, aerial and street imagery to identify property attributes for more than 9 million homes in Australia.
The approach allowed Suncorp to determine, for example, the size of a property, its distance to a body of water, or whether it has a swimming pool or solar panels – without having to ask customers complex questions.
“We saw geospatial technology as an opportunity to get better pricing without it being a new question to the customer,” says Steve.
“That was a key driver for us.”
The additional data simplified the buying experience, reduced the quote dropout rate and removed the need to verify property attributes at the time of claim, among other benefits.
However, when the system was first developed years ago, there were a lot of questions around how such an approach would work.
“There's a healthy dose of support and scepticism at that time,” Jim says.
“We decided to go forward and prove in a very short time that we can realise the value to the customer and to the business.
“In this particular use case, instead of asking questions from our customers, we wanted to go and understand all the properties in Australia and work out the factors that we use to rate them.
“And to do that, there's no right data set that we can use in the market. There's no government data sets available that show every property in Australia and how they are made, or what type of wall they have. So, we took a mammoth job on this because the only way you could do it is to take photos from satellites.”
During the initial phases, the team experimented with Google Street View and satellite data. They then began with small-scale testing before rolling out the approach across the company.
The team found that most of the time, the models are more accurate than the information provided by customers.
“We've done a lot of testing comparing data from customers versus what comes out of the aerial imagery and where they don't match. You know, we can look at the images and see who's right,” says Steve.
Such an analysis wouldn’t have been possible ten years ago because the technology to crunch the amount of data simply wasn’t there, says Jim.
“Google has a lot of this already in play and we used technology like AI and machine learning so we don’t have to manually look at all of it.
“Do they have a rooftop with solar panels or do they have a pool? It's a mammoth task, right?”
Advice for next year’s participants
For the Suncorp team, winning the Practice Prize is just the beginning. They’re already exploring other potential applications for the new approach, and they have a few ideas.
“There are use cases we want to go into because we have the data now. We still need to make sure that it's accurate, it's available and it's up to date,” says executive manager Nagender.
“So, my team is really working on that to establish that frequency of updates and trying to improve the efficiencies of our models to understand the data."
Emily adds: “We'll continue to explore the use cases that we have for this this amazing data set we now have. In the future, there could be possibilities to even extend to the motor insurance space to look at how this information is affecting a customer's driving behaviour as well as routes that they take on a regular basis.”
Nagender also has a word of advice for next year’s Practice Prize participants.
“Never put yourself down, just put up your hand, show what you have done. It's always good to see and to get recognised, whether you win it or not.
“We've never thought this project will win any awards when we started it, and we never thought it would get here. But that's how everyone starts.
“Just showcase your talent, focus on your innovation, and inspire others in the industry to do more for their customers.”
Applications for this year’s Practice Prize are now open, with submissions being sought from teams in the business, government, healthcare, education and non-profit sectors who have successfully implemented an innovative analytics solution in the past three years. The deadline for submissions is Friday, 24 March.
Finalists will be invited to present their solution at the 2023 Melbourne Business Analytics Conference on Wednesday, 2 August.
To learn more and apply, visit the Practice Prize page.