A prediction: You’re going to hear the term ‘prescriptive analytics’ a lot more often.
Described as the "final frontier" of analytics and the field's "third and final phase", according to Gartner 35 per cent of organisations will use some form of it by 2020.
The software market for prescriptive analytics is forecast to reach US$1.1bn by 2019, a 22 per cent CAGR from 2014, Gartner says. Market research firm Infinium reckons it will grow to US$5bn by 2021.
Figures like that have attracted a small but growing number of vendors to the burgeoning market, from global giants like SAS and IBM (which first coined the term prescriptive analytics) that are building out existing analytics offerings, to smaller sector specific players like Ayata (which trademarked the term in 2011), Earnix and MathWorks.
An Australian company has now entered the fray. Sydney-based Veriluma, which listed on the ASX via a backdoor listing in September, plans to break through this year by tackling a string of headline issues from the threat of a global pandemic to the the crisis facing the family law courts. Extreme weather disasters, dodgy financial advisers and domestic violence are also being targeted, the more high-profile the issue the better.
"That's the game-plan," explains its ambitious CEO Elizabeth Whitelock, "find the things that have a lot of focus and see if we can address them."
Informer
While predictive analytics looks at historical data to determine the probable outcome of a future event or chance it will occur, prescriptive analytics anticipates not only what will happen and when but attempts to explain why. It also suggests – or ‘prescribes’ – the optimal decisions to take and the potential implications.
Predictive analytics, says Whitelock, has failed us.
“It’s using data we’ve already gathered to try and predict what might happen tomorrow. Like driving a car by looking in the rear view mirror,” the Scot explains, “You don’t need to look very far in the world to see that in doing that we have failed to predict a number of different scenarios including Brexit and the Trump situation.”
Veriluma was born out of a Commonwealth Research Centre project for the Defence Intelligence Organisation in 2010. At the core of its patented engine is an algorithm based on subjective logic, which allows for analysing situations involving uncertainty and incomplete knowledge. It delivers a probability of a future event, and a confidence score that it will happen.
“It also allows us to take into consideration opinions and observations, the subjective, and recognise uncertainty. I don’t have a clue what’s around the corner, but I know someone who knows the area really well – so I can ask their opinion. That’s very different from the world of data analytics, which is all about gathering data and trying to make sense of what’s in it,” Whitelock says.
The ability to account for gut feeling and the unknown saw Veriluma run a proof-of-concept with the Department of Defence in 2012 to assess the reliability of intelligence supplied by informants.
The software allowed the department to account for an informant’s agenda, their bias and past reliability. The pilot successful, Defence continues to hold perpetual licenses for Veriluma software. To boost its efforts in national intelligence the company appointed former NSW Commander of Counter Terrorism Intelligence and National Counter Terrorism Committee adviser Mark Carrick to its board late last year.
The technology proven, the company is now turning its attention to major issues that it believes its software can solve.
“The great thing about it is it can address many different problems and industries. Given we’re a small company, we’re focusing on the things where there’s traction,” Whitelock says.
The big issues
One emotive issue in the company's cross-hairs is Australia’s ‘constipated’ family law courts. With waiting times for hearings stretching to more than 18 months, Veriluma believes its prescriptive capabilities can aid individuals and ease the strain on the system.
It formed a partnership with Legal Logix last year to work on a consumer-facing app.
“There are a large proportion of people who represent themselves in court. There’s a large proportion that do not get legal aid. How do we help them?” Whitelock says.
The app asks a long list of questions – both quantitative like the estimated value of assets, and qualitative like the reasons for a relationship breakdown – to generate an assessment of an individual’s likely chance of success and settlement in a dispute case.
“So they walk in knowing what they’re up for, if they understand what they’re entitled to, we can prepare a proper brief for them. And hopefully that means they can get through the system a bit faster, with a bit more confidence,” Whitelock says.
Eventually the software will be used to assess personal injury claims, estate litigation and immigration claims, Whitelock added. A separate pilot in the legal sphere was recently announced with corporate law firm Gilbert+Tobin.
Then there's the not minor matter of responding to climate disasters.
In December the company signed a six month contract with Defence to build a prototype ‘Indicators and Warnings’ model for Australia’s climate and disaster preparedness effort.
The project will involve 12 agencies, CSIRO and the Department of Foreign Affairs and Trade among them, to better resource the nation’s response to extreme climate events.
“We want to get to a place where we can see ahead of time what’s likely to happen and be responsive and proactive. To deal with things before the need for a big Defence operation which is a very costly exercise and should be a last resort,” Whitelock explains.
Previously the software had been used to build a model around a global pandemic. The recommendations for the government’s response “went to the prime minister and a pile of measures were put in place”, Whitelock adds.
Regtech effort SAMI, meanwhile, is a prototype designed to monitor and police financial advice providers.
“When somebody is going rogue… let’s avoid the media scandals and consumers losing trust, and create a mechanism whereby people end up knowing the advice has been given in a way that is in their interest, not in the interest of the adviser pocketing commission,” Whitelock says.
My two cents
The media-friendly use cases will help promote the value of prescriptive analytics, which Veriluma currently provides via SaaS, on-premise and headless, white-labelled offerings.
The goal is for Veriluma to eventually become an ‘adviser service’, its engine running in the background and its analytics engine built-in to other solutions.
“So we give you some advice based on what you’ve provided, and every piece of advice we dispense we take a clip of the ticket,” Whitelock explains.
Demand for prescriptive analytics and the guidance it gives users, is going to be huge she says: “One day we’ll have so many people having built applications that can run pieces of advice and we end up with a Twilio model that is two cents per piece of advice, because we’ll be doing millions of them. There’s my dream.”