Algorithms very well may be the future of marketing, as well as life in general. These powerful and mysterious beings are all-knowing, all-seeing and determine everything from Google search results to Netflix recommendations to driverless Uber cars. And they’re getting stronger every day.
I, for one, would like to welcome our new robot overlords. Their control of our every hope, thought and desire has really ushered in a new era of peace and prosperity for all. Long may they reign! /sarcasm
In all honesty, algorithms aren’t really anything to be afraid of. Except when they are. What does that mean for marketing? And what can you do about it? Can you really compete with an algorithm and outsmart it? Let’s explore.
Understand an Algorithm Is Just a Formula
Algorithm is a big fancy word dripping with connotations. All it really means is “formula.” An algorithm is nothing more than a millennia-old way of representing something equals something else. That’s it.
For example, when you enter “marketing advice” into a Google search, that term runs through algorithm that powers the search in such a way that your results equal “marketing advice.” At the end of the day, it’s really just a simple “if this, then that” statement.
Granted, they’re very complex and immersive statements, with a wide range of variables considered in any one instance, and the ability to process computations in microseconds, not hours. Still, an algorithm just tells you one thing equals another thing. Understanding this is the first step to thinking like an algorithm.
Algorithms Need Data to Survive
Since an algorithm is just a formula, you need information to put into that formula in order to make it work. No data, no working algorithm. They’re very primitive like that.
Going back to the above example, if there were no data inputs for the “marketing advice” search, there would be no results. The formula would quite literally not be able to function, silencing the algorithm. Yet you produce data every time you log onto the internet (or purchase something in a store, or exist as a person in the 21st century), so there really is no shortage there. Hence marketing’s emphasis on “Big Data.”
Each algorithm will use different data to produce the results it needs. Some rely on demographics and purchase history. Some will use context specific parameters. All of them need to define which data are important (and how important each datapoint is) before they can really get going. If you can break down the formula and understand its components, you can think like an algorithm.
Data Helps an Algorithm Improve Each Time
You’d call it learning. The algorithm calls it improving (or machine learning, actually). Remember how everything is data? The results of the algorithm are also data. What you do with those results becomes data that informs the next time the process is run. Confused yet?
In the simplest terms, this is how Amazon’s “Customers Also Bought” function works. Each cart that goes through to purchase become a data point. The algorithm then collects data on which products are frequently bundled together. Then, once it determines those products that are bundled together in purchases frequently, it can then suggest new things to add to your cart, entirely based on what other people have bought. It may also factor in other data such as user purchase frequency, geographic information based on your billing address, your personal purchase history and other demographics.
The point is the more people who purchase something in this case, the more data from which the algorithm can pull and the better the results would be. Then, the algorithm self optimizes to produce even better results for the next person next time. Maybe different factors need to be weighted differently. Maybe this product needs to be bumped down based on results. This is how the machine learns.
Think Like An Algorithm By Thinking Logically
No algorithm is infallible. Every single algorithm on the planet was either written by a person or discovered by one. None is perfect in every situation. There are flaws in every one of them.
That means the logic behind even the most complex algorithms were developed by man, and created to be as simple as possible so that man could understand it on some level. Thinking logically will help you think like an algorithm, using the components you’ve identified as important as the baseline of your formula. This is why you need to understand what information goes into the algorithm in order to outsmart it.
Of course, one of the gifts of modern man is the ability to think creatively, not just logically. Understanding the formulaic thought process of an algorithm is the first step. To get a few steps ahead of it, you need to figure out what the algorithm is missing. What are the blind spots? How can you improve its service? Where does it fall short? Fill in the blanks and you’ll have created a product or service that will beat the best machines out there, John Henry style.
Understanding the logic yet going above and beyond using your human brain. That’s how you beat the algorithms coming to take your job. Who knew?