COMMENT | What’s one thing you could study that would impress every customer you talk to (apart from Italian)? What’s one thing that will affect every technology you sell? No, it’s not electricity. What’s one thing that hardly anyone knows anything about, but will reshape society? I’m not talking about Trump’s direct messages on Twitter.
Algorithms are everywhere, and yet, they’re largely invisible. We hear all the time about Facebook’s algorithms and Google’s algorithms – and we should, as these logic processes influence the behaviour of literally billions of people. But there are algorithms at work in business software too, and how they work and what they do are crucially important to your customers. The more you know about algorithms, the better you can use the software or service they underpin.
The most common business algorithms are in marketing; it’s how all of those Facebook and Google ad campaigns are finding potential new customers. When you start playing with custom audiences, placing pixels on your website and uploading your email database to Facebook, you’re trying to train an algorithm to find the perfect prospects.
This can be a difficult concept for people to understand. Business owners want to see results rather than discuss logic flows. But if you know the flow of logic behind an algorithm, you can find combinations that will unlock lucrative pots of new customers.
Take Facebook as an example. If you hold an event and get a heap of leads to follow up, they can form a custom audience which Facebook will try to match.
A business owner or marketing manager who understands the concept of custom audiences could cross-reference the events list with a list of customers from their accounting software, filtered by those that spent $100,000 or more last year and live in Victoria. The resulting smaller and more targeted list would yield a very high-value group of prospects. The business could then send highly attractive discounts to the best kind of new customer.
It’s often impossible to know the exact logic that an algorithm will follow. However, you can think about the consequences of the algorithm’s decisions, and how those consequences will affect a company.
CRMs also use algorithms to assign deals to sales staff with spare capacity. A business may find that its best sales staff are always the busiest, and all new leads are going to the least trained salespeople.
The algorithm won’t necessarily know which sales staff close more deals. It won’t think about the consequences of giving the greenest staff member the responsibility of creating a good first impression. It’s better if the business owner has that knowledge so they can decide if it’s what they want.
Other common algorithms are route planning in job management software, restocking orders in inventory apps and transaction coding in accounting software.
If you want to distinguish yourself from others in the field, talk to the software vendors about how their algorithms think. Few people are doing this – inside or outside of the channel – which is worrying given how much algorithms control our lives.
There is a movement for algorithmic transparency, so that people can find out how decisions are being made and check the ingoing data or the outcomes for discriminatory or ethical issues.
“The core problem with algorithm-based decision-making is the lack of accountability. Machines have become literal black boxes – even the developers and operators don’t
fully understand how outputs are produced,” said Marc Rotenberg, executive director of the Electronic Privacy Information Center, in a February 2017 research report by the
US-based Pew Research Center.
“We need to confront the reality that power and authority are moving from people to machines. That is why #AlgorithmicTransparency is one of the great challenges of our era.”
It’s a challenge that must be met, but it’s also a business opportunity for those who move on it quickly.
Sholto Macpherson is a journalist and commentator who covers emerging technology in cloud