Automation is the greatest change we’re going to see in business over the next decade as companies increasingly rely upon computers to make day to day decisions.
Giving control to algorithms however comes with a set of risks which managers and business owners have to prepare for.
Earlier this week the risks in relying on algorithms were shown when car service Uber’s management was slow to react to a situation where its formulas risked a PR disaster.
Uber’s misstep in Sydney shows the weaknesses in the automated business model as its algorithm detected people clamouring for rides out of the city and applied ‘surge pricing’.
Surge pricing is applied when Uber’s system sees high demand – typically around events like New Year’s Eve – although the company has previously been criticised for alleged profiteering during emergencies like Hurricane Sandy in New York.
In the light of previous criticism, it’s surprising that Uber stumbled in Sydney during the hostage crisis. Shortly after criticism of the surge pricing arose on the internet, the company’s Sydney social media manager sent out a standard defence of surge pricing.
We are all concerned with events in CBD. Fares have increased to encourage more drivers to come online & pick up passengers in the area.
— Uber Sydney (@Uber_Sydney) December 15, 2014
That message was consistent with both Uber’s business model and how the algorithm that determines the company’s fares works; however it was a potential disaster for the business’ already battered reputation.
An hour later the company’s management had realised their mistake and announced that rides out of Sydney’s Central Business District would be free.
Uber rides out of the CBD today are free for all riders to help Sydneysiders get home safely. See http://t.co/UIwoom25Bm for more info. — Uber Sydney (@Uber_Sydney) December 15, 2014
User’s mistake is a classic example of the dangers of relying solely on an algorithm to determine business decisions; while things will work fine during the normal course of business, there will always be edge cases that create perverse results.
While machines are efficient; they lack context, judgement and compassion which exposes those who rely solely upon them to unforeseen risks.
As the Internet of Things rolls out, systems will be deployed where responses will be based upon the rules of predetermined formulas.
Businesses with overly strict rules and no provision for management intervention in extreme circumstances will find themselves, like Uber, at the mercy of their machines. Staking everything on those machines could turn out to be the riskiest strategy of all.