BillMonitor: Mobile contract comparisons, now with science
Oxford eggheads have built a Web site they say can save you money, by analysing your online bills using mathematical mojo
We always suspected it takes a PhD in mathematics to understand mobile phone contracts. Now Oxford boffins have proven it takes three of them.
Maths researcher Stelios Koundouros and two colleagues decided to apply their planet-sized brains to mobile tariffs after he couldn't figure out his own contract. They created BillMonitor, a new tariff-comparison Web site, to find order in chaos.
Like other comparison sites, such as Mobile Phone Checker and moneysupermarket.com, you can tell it the number of minutes and texts you want and it'll recommend a plan from its list of 46,000 tariffs. But if you're willing to let it digest your online bills, things get really impressive.
We gave the site our usernames and passwords so it could grab our billing data for the past three months. About a minute later, after some heavyweight statistics voodoo, it emailed us links to a list of contracts that would save us money.
Some of us won, potentially dropping £20 a month from our bills. But the system falls down when it can't reach your online statement, or if you're on an old tariff that can't be beaten by anything the networks are currently offering.
We especially liked the groovy graphage, which shows a breakdown of your data, roaming, texts and calls. Even if you're not thinking of switching, BillMonitor is great for dissecting your mobile usage in a way that's much clearer than most networks' bills.
The recommendations are based on anonymised data from 100,000 users provided by an equally anonymous European network, as well as the data of everyone who's used the site.
Your usage is combined with data from similar users and put through the statistics blender to predict which plan would be cheapest in the future. By using a hefty dataset, the system isn't thrown by unusual usage -- like a heap of roaming charges while on a once-in-lifetime holiday, for instance.
But no statistical forecast is perfect, so the system is cautious. It'll spit out a contract that's very likely to save you a little bit, rather than one where you could save more, but possibly lose you money if the forecast is off.
It's all still in beta, as the developers wrestle with automatically reading online bills that don't have a standard format. They also have to manually update the contracts with any special offers and bundles on offer.
Like a statistics superhero, Koundouros is happy to be using his powers for good. "We want to use advanced statistics to help people," he said.