Technology

What is an algorithm? An explainer

Very likely you’ve taken part in your very own algorithm.

By Lachlan Brown

15 December 2021 · 3 min read

An algorithm is a set of step-by-step instructions that is used to complete a task.

When you last did your washing, did you sort your clothes?

One pile for lights/whites and another for darks/colours?

If so, you’ve taken part in your very own algorithm.

Put simply, an algorithm is a set of step-by-step instructions that is used to do something.

As that is a very broad description, you can imagine algorithms surround us, in digital and ‘analogue’ life.

In most cases, you’ll be experiencing the work of algorithms when using your computer or phone. This is because a piece of software can follow a step-by-step algorithmic guide much more quickly, and precisely than a human.

Trading on the stock market was traditionally a human’s role, but these days, computers typically account for 50% to 60% of market trades, according to Art Hogan, chief market strategist for B. Riley FBR. When the markets are extremely volatile, they can make up 90%of trades.

When we break down algorithmic trading, we get a good understanding of what a digital algorithm involves.

For trading stocks, an algorithm will execute a large order on a financial market using pre-programmed instructions accounting for variables such as time, price and volume. These instructions might include:

1. Buy 50 shares of a stock when its 50-day moving average goes above the 200-day moving average.
2. Sell shares of the stock when its 50-day moving average goes below the 200-day moving average.

Algorithmic trading has also been extremely cost-effective for financial institutions, given software can follow explicit instructions, monitor prices and place orders endlessly, without hankering after a salary or bonus.

Today, digital algorithms like this are ingrained in everyday life. So much so, that they are beginning to shape the way we live day-to-day. Things like marking students' homework, creating original artwork, making national security decisions and writing legal documents are all tasks that algorithms undertake on a daily basis.

Interestingly, the music industry uses algorithms to find new artists. They know the formula behind the best pop hooks and can go out and search for any undiscovered track that matches the winning recipe.

So that new number 1 hit you were listening to on the radio, may not have been discovered at the humble open mic night, but by a very sophisticated robot on the web.

One day even doctors will become algorithms. There is already a fully automated pharmacy at the University of California in San Francisco. It has doled out 2 million prescriptions without making a single mistake, an average human pharmacist would make 20,000 mistakes with the same amount of prescriptions.

Thankfully not all algorithms are created to replace a human’s job. There are cases where they have been implemented to make some jobs easier and more efficient.

Mattersight, a company based in the US, uses an algorithm to determine the personality type of an inbound customer service caller, then directs it to the most appropriate customer service rep.

When you call up a company that uses Mattersight, you’re automatically transferred to an agent who can best satisfy your needs based on your personality and the company representative’s past performance and personal strengths.

The moment you start talking, your tone, tempo, keywords, grammar, and syntax are all fed into an algorithm that determines which agent would be best suited to handle your issue.

Someone who calls up hysterical, wanting to know why their flight has been cancelled, would likely be transferred to a different agent than someone who casually asks what movies are available on the flight.

If you were to call up and say something along the lines of, “I’d appreciate it if you'd help me to understand,” you would be treated differently than someone who was to call up and say, “I’ve got a problem — you need to fix it.” It’s not just a matter of who’s politer — it’s a question of how people articulate their needs and prefer to receive information.

Mattersight has listened to over 1 billion conversations and created a library of 6 million algorithms to categorise the human language. They have effectively built a bot that will likely understand your personality better than you.

Mattersight products are based on a human behaviour tool called the Process Communication Model. The model was developed by Taibi Kahler, Ph.D., and has been used for years to select astronauts at NASA, among other things.

The effect of this technology results in the calls being half as long and they come to happy resolutions 90% of the time instead of 47% of the time.

So not only are algorithms being used to write prescriptions, find the songs we hear on the radio and trade on Wall St, but they are now finding out what kind of people we are and determining who we are best suited to talk to.

The future of algorithms is uncertain, however, it is fair to say that they’ll be influencing our lives for many years to come.