Lifting the lid on personalised pricing

Australian banks and insurers will have to navigate privacy, consumer law and public sentiment as they move to individually price products using big data analytics.

Organisations including IAG and the Commonwealth Bank are looking to combine social media with customer data to tailor insurance, loans and credit card products to individuals.

Resultant personalised pricing schemes, which IAG expects to deliver in two to three years, will likely benefit healthy, ‘low-risk’ individuals with steady incomes.

For people on the other end of the spectrum, however – those statistically more likely to make insurance claims or default on loans – personalised pricing may shut doors.

“This is an area that is likely to get increasing policy focus and attention,” according to Gilbert and Tobin partner Peter Leonard.

“Individual pricing has a real risk of creating isolated communities or groups that are effectively uninsurable, or can only obtain insurance at great cost, and might be unable to obtain financial products.”

The British Office of Fair Trading in November launched an investigation into personalised pricing, to address concerns that businesses were using individuals’ web browsing data to modify their prices.

But aside from the health insurance sector, there appear to be no Australian laws against personalised pricing.

Australian health insurance is “community rated” under the Private Health Insurance Act, so firms must offer the same price to all customers regardless of health status, age or claims history.

Leonard expected big data and individual pricing to be debated increasingly in Australia as companies and regulators worked out “what people can do and what people should be allowed to do”.

While he said there was nothing in Australia’s privacy, competition and consumer laws to restrict individual pricing, Leonard warned that companies could run foul of discrimination law if prices were affected by ethnicity or religion: characteristics easily determined through social media.

Several insurance industry groups approached by iTnews declined to be interviewed on individual pricing as they said they were unfamiliar with the technology.

Technologists, meanwhile, were wary of being drawn into discussing privacy and regulatory issues, stating only that their organisations met all regulatory obligations.

One prominent Australian big data expert suggested that more advanced analytics work was going on behind the scenes of the local finance sector than the public realised.

A CommBank spokesman acknowledged that it was trialing big data technology but declined to address questions on personalised pricing.

The marketing angle

To date, much of the public discussion about big data has pertained to its use in marketing and identifying cross-sell opportunities.

Westpac’s head of customer relationship management and digital Karen Ganschow is leading an initiative that analyses customer behaviour to offer a product, service or tip.

Ganschow told an AIIA panel last week that the program used transaction and behavioural data, such as clickstream data, to identify “Next Best Offers” for one in five customers.

“[Branch or call centre staff] can see that the customer has got a Next Best Offer against them,” she said.

“We try to put in the script some nuggets of gold about why this customer would be interested, what’s the profile we know about them and what are the insights we have.

“The key thing is relevance. We’re a big believer in the Net Promoter Score … when the banker has made a pro-active recommendation, even if a customer doesn’t accept it, the fact that they can see that it is personalised for them drives up the NPS for that interaction.”

Ganschow said later phases of the program would draw in data from social networks to provide more insight into life events that were typically “critical financial moments” for individuals.

Westpac hopes to generate Next Best Offers for 80 percent of customer interactions by 2018 and display the information on new channels, including online banking.

According to IBM Global Business Services Business Analytics partner Amit Bansal, the rise of real-time analytics has allowed businesses to consider customers at a more granular level.

“Traditionally, all organisations have looked at large demographics, putting us all into buckets of ages and lifestyles then marketing to us,” Bansal said.

“Now, with all the data available, you actually know more about the individual and every individual is different. That’s where the big shift is coming.

“The old school thing in analytics was ‘tell me what questions you’re trying to answer and I’ll go and find the data and model it for you and come back to you in six months’. The problem with that is customers have moved on six months later.

“Technologies have changed. We can look at all the unstructured information and all the bank’s data and let the analysts figure out what questions they should be asking.”

Read on to find out how insurers are cutting costs by identifying health risks and what's holding Australia back from personalised pricing.

Improving health outcomes

While health insurance pricing is bound by Australia’s community rating system, insurers may tailor packages to individuals, or even use analytics to improve their health.

HCF’s My Health Guardian program is one example. The program is linked with health insurance claims and membership data to identify and proactively advise members of risk of hospitalisation.

HCF chief information officer Patrick Shearman told iTnews in December that My Health significantly reduced the number of members requiring hospitalisation.

My Health Guardian has about 22,000 users and targets issues such as weight loss, smoking, blood pressure and cholesterol.

Bansal said overseas insurers had implemented similar technologies, citing one income protection insurance provider that had reduced the occurrence of depression among injured white-collar workers.

“We took unstructured content and claims information from their call centre logs and claims... and we were able to understand and predict that a white collar worker with a back injury has a very high propensity, six months down the track, to be depressed.

“By analysing the data you can say, ‘you know what, I need to find those correlations so I can intervene earlier to prevent that happening’, so you’re not only preventing [cost] for the organisation, but you’re also giving good service to the customer.”

Another US insurer discovered that customers in the nursing profession had the highest propensity of breast cancer-related claims.

As a consequence, that insurer began proactively educating customers in that demographic of breast cancer and the trend it had detected, Bansal said.

The data debate

But aside from the handful of industry standouts, most organisations appear to be in the early stages of big data analytics, with many still struggling to pull together information silos.

Suncorp has moved to decommission its six enterprise data warehouses in favour of a platform that scrapes data from various backend systems and allows it to be analysed.

“I think one of the challenges most banks face is thinking about analytics and information management as very closely related but different aspects,” said IBM Global Business Services partner and financial services sector leader Kevin Jury.

“Whilst you can drive analytics in a big way and do lots of interesting things, it really requires a key ingredient: the data.”

“I think it’s going to be a journey of ‘grow the information management platform’,” Jury said.

“[Businesses will] put the data on the launchpad, do some analytics work … and as more devices get deployed out there, be they sensors or whatever they are, you’ll see more analytics stringing together that sensor data, coupled with mobile data. Then, once again, you’ll end up with a cycle of getting more data up on the launchpad, to do more analytics.”

Last July, iTnews reported that Australian banks were lagging behind their international counterparts in using social media outside of the usual communications and customer support areas.

Ultimately, Gilbert and Tobin’s Leonard suggested that public sentiment would hold the most sway over businesses’ use of data.

He called for an “open and informed public debate” on big data that would raise not only regulatory boundaries but also the risks and benefits to society.

“When you actually look at some of the positive applications of big data today, they are truly amazing, and yet the dark side of big data is also undeniable,” he said.

“I advise people that there are two things that you always need to consider with big data: whether you can do it and whether you should do it. In the second consideration, you apply amongst other things the ‘spookiness’ factor.

“Usually the issue stopping what you might do is concern that the individual might regard that as an unreasonable intrusion upon their privacy.

“I think most banks and insurers are very conscious of their public perception and reputation and are conscious that if they use big data to its fullest, it may lead to results that are not acceptable to consumers and will lead to regulation.”