BBC News

How algorithms will help us spend, spend, spend

By Dave Lee
Technology reporter, BBC News

image captionEver feel like someone is trying to tell you something? Sales are already tempting - but with new algorithmic technology, buying may become even harder to resist

We've all heard it. That little imaginary companion, sitting on your shoulder. "Go on," it says. "You deserve a treat. Buy it."

That "it" could be anything. Clothes, shoes, gadgets - we all have our vices.

But what if that imaginary voice - which is, willpower permitting, under your control - one day became real?

DBS Bank calls it a "personal concierge", and it's best understood by picturing yourself in an expensive designer clothes shop.

Your smartphone knows where you are - thanks to the GPS location technology found in many apps these days - and it alerts your bank through an automated system that you've signed up to.

As well as knowing you've got a history of buying from similar stores, your bank also knows that you're running a bit low on cash at the moment.

It realises the chances of you buying may be pretty low. That "voice" on your shoulder is pretty quiet as realism sets in.

Ordinarily, unless you wanted to go hungry that month, you'd leave the store.

But wait! Your phone beeps. A text message. Buy it in the next 20 minutes, it could say, and you can borrow the money at a good rate. Not only that, but you'll get 20% off the clothes.

You leave the shop, bags in hand. You've been a successful participant in a process known as "right place, right time" marketing.

'The new slave'

Of course, even banks like DBS, one of Asia's biggest, can't afford to pay an individual to keep tabs on all of its banking customers, and so it adopts technology which is growing rapidly in many industries: intelligent algorithms.

"Algorithms are effectively the new slave," explains John Bates, chief technology officer for Progress, a company which specialises in creating algorithm platforms for a wide range of businesses.

He says that in an ideal world, banks - or any loyalty-dependent business - would be able to offer each and every customer a personal assistant keeping an eye on them so they could target the ultimate offers to suit the individual. But logistically, of course, that's out of the question.

"They can't afford that - not even by outsourcing to India or China or wherever the cheapest place these days.

"So you're going to use technology. Algorithms don't need to go to the bathroom, algorithms don't need a pay cheque."

Progress works with a number of businesses who use algorithms to make key business decisions.

Not all of them are prepared to go public - the trade can sometimes seem like a shady business, guilty of pigeonholing consumers into pre-determined groupings. Others argue that such systems could even be racially profiling people.

But the techniques are gathering momentum, lead by evidence of rapidly growing profits and effectiveness.

Turkcell, Turkey's largest mobile telecoms provider, recently won an award for its implementation of algorithms. Similar to the efforts of DBS Bank, Turkcell - also a Progress client - was targeting offers based on a user's location.

"We have seen that those offers sent using the real-time marketing system result in up to ten times more positive responses than those offers sent using conventional methods," says Turkcell's chief consumer business officer Emre Sayin.

Automatic investments

Elsewhere in the business world, algorithms are being used to tempt people to splash out on a lot more than just luxury goods.

Venture capitalists like David Coats, are using it to make crucial business decisions.

His company, Correlation Ventures, invests in technology start-ups based on various criteria that is crunched by its own algorithm.

If a start-up passes the test, it earns investment.

image captionDavid Coats, second from right, has 20 years' experience of investing in companies

"What we started to notice was how time-consuming and distracting the fundraising process often is for entrepreneurs," Mr Coats told the BBC.

"All that time taking away from time they could be spending managing their business."

Start-ups looking for investment will submit around five documents detailing financials and other business information.

This data is fed into a program which compares it against a database of over 20 years of venture capital activity.

Those who pass the algorithm test will be given a 30-minute interview, go through some background checks and then... get investment. It all takes about a fortnight.

"We do not repeat classic due diligence," Mr Coats continued.

"We commit to how much money we're going to invest in the company over the company's life, and we do not take board seats."

Gut feeling

While this approach may appear a little gung-ho to hardened business people more used to trusting their gut instinct rather than a computer, Mr Coats said the system is designed to complement existing systems.

"That gut feeling, and the lead-VC diligence is vital," Mr Coats insists.

"This is not replacing that, it's an overlay."

Crucially, Correlation Ventures is never the first investor in a company. Before firms get to Mr Coats and his team, they will have been backed by other investors who have spent lengthy time checking out their credentials.

Their scheme, Mr Coats said, is simply about shortening the time it takes to get more money on board.

Correlation Ventures is looking to build a portfolio of between 75-100 companies. So far, it has backed 26 - with a new one turning up at a rate of roughly once a month. Much quicker than your typical investment firm, they said.

"We're consistently getting accolades from entrepreneurs about how fast our decision making is, and how we're meeting an unmet need," Mr Coats said.

Other figures in the venture capitalist industry are said to be equally as impressed.

"The reaction has been tremendous," he added.

"When I founded it I wasn't sure what the reaction would be, but we've already made co-investments with many different venture funds."