Ignia and Microsoft Azure help start-up tackle global drive-away theft epidemic
It’s the genesis of most great innovations: first find a problem then work out how to solve it.
For Perth’s Eoin Byrne, the problem came to him as he was heading home with the rising sun after working security a night club. Eoin was involved in integrating Perths first ID scanners at licensed venues in town; essentially a quick history checks of a patron’s behavior at other venues.
He pulled into his local garage to get petrol and noticed the guy behind the counter scribbling down something as he looked out the service station window.
His interest piqued, he asked the attendant, “What are you doing, because every time I come into the petrol station you guys are madly scribbling stuff down on a notepad?” The attendant was doing what attendants do throughout Australia and much of the rest of the world; he was writing down the number plates of vehicle with suspicious looking occupants and trying to quickly scan through a massive list of hundreds of numbers to see if that car had previously been involved in fill-and-drive-off theft from the station.
“And that was the trigger for the idea, because Eoin thought, ‘Hang on a second! BP can drill to the bottom of the Bering sea for old dinosaur bones, but they can’t automate this? If only BP could do a quick history check on there customers too, now that’s an opportunity”
Eoin approached Anthony Schmidt about this. “Eoin came and talked to me about it because of my operational expertise and background. I said, ‘Oh mate, I think you’ve got something here!’”
The pair became partners and Scancam, innovative technology to solve the problem of fuel drive-offs, was born.
It is a problem worth solving. Estimates of the cost of fuel thefts in Australia vary; one official figure put the cost at $60 million, but with many thefts going unreported because of the difficulty of recovery, industry experts believe the real losses exceed $120 million.
After much research, Eoin and Anthony found that the initial problem of theft was exacerbated by a slow, manual reporting system that involved burning CCTV footage onto a disc, a written incident report and a trip to the police station to hand in evidence that rarely resulted in a conviction.
The pair decided that licence plate camera recognition was the starting point but soon discovered that even though the technology was in use for speed cameras there were extraordinary difficulties surrounding taking photographs in the busy, unpredictable environment of a service station forecourt.
Already aware that a perfected Scancam would have a world-wide market, they made the decision to only use the very best of equipment – “no shortcuts, no cost-cutting,” said Anthony.
They sourced one of the world’s most advanced cameras from Mobotix which, after the writing of special software, proved adept at licence plate recognition. By November 2014 they had an app and a prototype and went about proving that it all worked.
After a few months of successful trialing, Microsoft partner, Ignia entered the scene around the middle of 2015.
“Once we had a working prototype we then decided that we needed to build something now on an enterprise scalable level that will support multiple time zones, regions, countries, support 50,000 devices allowing us to build a big business using this platform,” Anthony said. “That’s where Ignia and Microsoft Azure came in.”
The Scancam Project started as a migration from an AWS hosted Ruby on Rails/PostgreSQL backend application to a Microsoft Azure backend (Web App, Queues, WebJobs and SQL Database).
“Midway thru the project, we decided to replace their Objective-C based iOS application as well and we started work on rebuilding the mobile app on Xamarin and Azure Mobile Services,” Anthony said.
The company has also recently added into the mix Raspberry Pis running Windows 10 IoT core for the TV display functionality with the intent of using these same devices for on-site monitoring and image collection.
Anthony said the Scancam team were attracted to Azure for a number of reasons – the level of support they received, the ability to scale up and scale down, the capacity to collect data and being able to incorporate other features.
“I mean, down the track we could be scanning over three million licence plates a day, so that’s going to become extremely valuable data for maybe Coles, for a loyalty program, or the police or anti-terror or homeland security or things like that,” he said.
“We’re also able to bolt on other features and other datasets, so, for example, the WebArt or Fuel Watch and things like that.”
Using Azure Machine Learning to analyse and predict patterns in drive off related to fuel prices, weather condition’s, demographic areas and more Scancam are looking to being able to predict a drive-off through facial recognition software combined with registration details.
The Internet of Things is also opening an even more exciting future, with plans to have a screen on every bowser which greets customers through their licence plate: ‘Welcome xyz123, your bowser is now ready for use’ or if the licence plate appears on a fail to pay list the screen could say: ‘Please pay in store before refuelling.’
The Scancam team envisage service stations being able to offer loyalty rewards for regular customers and selling media space on the screens.
“I’m sure there’s many more applications we can use it for, but its opening up more avenues for potential business streams and giving more substance and strength and capability to our existing platform,” Anthony said. “Because that’s a great selling point; the first question everyone asks is what happens if my internet drops out, and we say no problem, it continues to run, we’ve got it covered locally until you’re restored.”
Ignia Director Joshua Boys recalled the pitch to Scancam, “We went through what the platform offers in detail, focussing on its core features and scalability, and they were totally won over.”
The existing system was changed from Amazon to Microsoft Azure within two weeks. “Scancam love the platform and so now we’re just adding features,” Joshua said. “They were 100 per cent non-Microsoft and now they’ve totally converted.”“They’ve ported their websites, moving from Google to run their organisation on Office 365, with the whole app in Azure. We’ve managed to architect everything really efficiently to ensure they have the opportunity to scale, and we’re also porting them from iOS to Xamarin to provide consistency between their front-end and back-end coding.”
Incredibly, considering the complexity of the problem and the innovation involved in the solution, Scancam went live in October and is likely to launch in Britain early in the New Year.
The new Scancam backend system has been receiving plate recognition data for four sites over the last month (roughly 14.5k vehicles) and will soon migrate their first site to the new backend as well. Training on the new mobile app and the corresponding customer portal has started with the intent of finally deploying the mobile apps by end of October.
The problem has been tackled on two fronts – first as a deterrent and then as debt recovering from those not deterred.
“CCTV is no longer a deterrent,” Anthony said. “So we came up with the idea to actually display the licence plate. What we’ve got is a display screen up the front, just at the main doors, and it shows everybody’s licence plates that’s currently filling up at the bowser.
“The day we put that up the store didn’t experience a drive off for two weeks, which is unheard of – they never went two weeks without a drive-off.”
The software does a database check of known offenders, delivering the results almost instantly to a tablet which allows the attendant to keep the bowser closed and to ask the motorist to pay before filling up.
Scancam has also partnered with a major debt collection agency which will commence debt recovery proceedings, usually within 24 hours. Current indications are that 82 per cent of debts are recovered on the strength of a first letter of demand.
“Even if we get 25 per cent back the petrol stations are going to be doing well,” Anthony said. “In fact, they’ll be paying for the system, which will become cost neutral.”
Any reduction in the incidence of theft also helps the stations because much of the profit comes from incidental purchases within the service station store.
“We’ve got the deterrents, we’ve got the alert, the database checks, and now we’ve got the debt recovery,” Anthony said. “Put all those things together and we’ve got a very compelling business case, which means people are knocking down the door at the moment.
“It’s never ever been offered before to the retail industry in Australia and as far as we can see around the world, because there are massive drive-off problems in other markets globally, which is where we’ve set our sights on down the track.”
The company also recently took away a gong at the national iAwards for research and development. And the high profile of Scancam within the start-up community has people also knocking on Ignia’s door.
“We’ve received increased interest from other established start-ups with strong revenue streams who have heard about our work with Scancam. There are a number of companies who are yet to reach their full potential simply because they haven’t been able to achieve the right price-performance ratio from their technology mix,” Joshua said.
“With careful, clever architecture through Azure, you can get enormous economies of scale with little increase to costs.”
Scancam, Ignia and Microsoft Azure . . . solving problems throughout the world.