Inside Origin Energy’s skunkworks analytics program

How a company credit card and a free Wi-Fi hotspot were essential tools as Origin began exploring the potential of predictive analytics

Shadow IT has a bad reputation among people whose job it is to police an enterprise’s use of technology. But sometimes it can be worth it. Ask James Moor, for example.

Moor, group manager – market risk at Origin Energy, told the Amazon Web Services (AWS) Summit in Sydney that investing in a cloud-driven data and analytics capability had paid off for the company.

Getting to that point, however, involved a “certain amount of stealth” for Moor and his team, he told the conference.

“Step one for us was about executing a number of proof-of-concept projects,” he said.

“We needed to prove that there was a different way of doing things – that we could be agile, that we could be innovative and that there was a business case for doing so.”

However Moor’s team found itself unable to open its first AWS account through Origin’s procurement function because the company was not an approved IT provider.

“So we made a minor infringement of policy,” Moor said.

“Our first cloud account was run off my corporate credit card. And then later when we discovered that the corporate firewalls wouldn’t allow appropriate connectivity to the cloud. The initial solution, now the stuff of legend internally, was the free Wi-Fi from the hotel across the road.”

“We did quickly graduate to something more sensible: A wireless dongle from Officeworks,” he added.

It was a “blackops approach to getting things done”.

“In this way we were able to rapidly and very cheaply get some proof-of-concept projects up and use them to start forming a business case within the organisation,” Moor said.

Those projects allowed Moor to the team to pursue upfront investment in a number of key enablers to develop Origin's data and analytics strategy, including a data lake and a suite of analytics tools.

Those investments were embedded in a number of commercially viable projects, he added.

“So we were asking management to invest in a program of tangible commercial projects and our data and analytics enablers get a free ride,” Moor said.

Origin is a vertically integrated energy company that covers the entire gamut of operations from primary extraction through to power generation and retail.

It’s Australia’s largest energy retailer with more than 4 million customers, and it participates in more than 100 million customer interactions per annum.

“In this respect, Origin would look a lot like many other retailers,” Moor said. However the company is preparing for a tidal wave of data as basic meters that are manually read four times a year are progressively replaced with smart meters that have built in telemetry and report energy consumption every half an hour.

“This means that Origin’s consumption database jackpots from around 16 million data points per annum to over 70 billion,” Moor said.

That data tsunami is an exciting opportunity for Origin, Moor said.

“How are we going to leverage this data to bring new products to market and to enhance our customers’ experience?” he told the conference.

Origin’s analytics program has so far facilitated around half a dozen initiatives, including its ‘predicable plan’ offering that gives customers the option of a fixed total cost energy plan and is designed to eliminate bill shock.

Another key outcome has been a gas availability project, which centres on making sure that Origin has enough gas available to its approximately 1 million natural gas customers

“This turns out to be a highly computational intensive problem,” Moor said.

“Demand is inherently uncertain – it depends predominantly on weather – and gas supply can come from many sources and each source has complex characteristics of how much gas for how long and under what circumstances.

“Finally, supply and demand are linked by a complex network of transport and storage options each with unique capacity characteristics.”

The company’s previous solution, which relied on local hardware, was “more or less broken,” Moor said. A full simulation would take around four days.

“With the flexibility and agility afforded us in the cloud, we were able to experiment, to innovate and to rebuild things, so the equivalent simulation now runs in under 10 minutes,” Moor said.

“Now this is a game-changer in terms of business value: Not only is our analysis now more timely and relevant, but the reduced run time means we are able to test far more scenarios, thereby providing a much greater range of insights than had previously been available.”

An additional pay off was cost savings. The previous solution required a $250,000 investment per year in hardware. The cost has now been slashed to $60 per simulation — about $2000 per year — with zero upfront costs.

For Moor, at the heart of Origin’s data journey is an understanding that developing an analytics capability is linked to developing a culture of innovation.

“You don’t build a data and analytics capability cognisant of exactly what innovation someone will come up with – what new products someone will invent and take to market or what new initiative or insight someone will uncover to enhance your customers’ experience,” Moor said.

“Rather, the destination we are targeting and the point of this whole strategy is about building a platform that will enable a capability and will promote a culture of experimentation and innovation.

“And it’s the capability and the culture that we are seeking that will ultimately deliver you your yet-unknowable commercial outcomes.”

Join the Computerworld newsletter!

Error: Please check your email address.

Tags data analyticspredictive analyticsAmazon Web ServicesdataCloudbig dataOrigin EnergyAmazon Web Services (AWS)cloud computinganalyticsOrigin

More about Amazon Web ServicesAWSOfficeworksOriginOrigin Energy

Show Comments

Market Place