Two things: the growing field of Optimization and the age-old concept of Network Effects.
Why does this help explain why these platforms can be so addictive? It’s because the companies running these platforms know that they are part of the attention economy. If they take their foot off the pedal, our attention to wander to one of their competitors. Netflix’s CEO, Reed Hastings, has said that its main competitors are Facebook, Youtube and sleep. Our attention on is the economic engine that enables a lot of these tech companies to thrive.
So what do these companies do to encourage us to devote as much of our attention to them as possible? They test. They test and test and test and test. Facebook is running more tests in a single day than the FDA runs in a year, according to Seth Stephens-Davidowitz, author of the book Everybody Lies.
The tests they run are called A/B tests, or simply referred to as optimization. What companies like Facebook do, is they develop teams of experts that understand their users. These experts develop hypotheses about what will make the user experience better, or more “sticky”. These hypotheses can be anything from changing the color of a button to drive higher conversion rates, to radical changes to its interface to increase the amount of time users spend on the platform.
Then the hypothesis is tested in the form of an A/B test. Even though its called an A/B test, it can assess any number of variables. The button color test can test eight different colors, but it is still an A/B test. Multivariate testing comes into play when the impact of multiple variables are being tested in combination – maybe the new interface combined with the button color change will result in the highest amount of conversion.
Crucial to the A/B test is that “winning criteria” are established. This is means that the testers define metrics (conversion rates, time spent on site, clicks, etc.) whereby if the test variable can show a statistically significant increase in that metric, then the test is considered to be a winner. If a statistically significant winner cannot be found, it’s back to the drawing board.
To test these variables, identical groups of users are shown the control (the way the button looked before) and the test (the new button color). The test runs until the size of each group reaches a statistically significant sample size. Usually this means that testers believe that the result they see will be repeated 95% of the time.
What happens when a winner is declared? The change is likely implemented on a broad scale across the site. But the testing isn’t done there. Since these companies have very powerful and sophisticated software and algorithms, they are able to compound successful tests to determine the optimal state of their website. If a blue button drives more clicks than a red button, maybe it’s time to test which shade of blue drives the most clicks. Social media platforms are able to determine what experience, content, colors, font size drives the most usage by specific groups of people. I may see a royal blue button, my wife might see aquamarine.
This is by no means exclusive to social media companies. All companies with an app or website do this. It is often said that no two Amazon websites look alike. Netflix does this as well, constantly testing new interfaces and ways to recommend shows to us. However, I am deeply skeptical of Netflix’s own valuation of their recommendation algorithm at $1bn.
This is why it so difficult for us to put down our phones and avoid social media. Every ounce of our will power is up against teams of experts that are constantly running tests to determine how to make it harder for us to put the phone down. Apple appears willing to aid in the fight against our smartphone addictions, but it will take more than an app that tells us how much we use each app to cure our societal ills.
This brings us to network effects, a simple concept, that can help us understand why social media platforms such as Facebook and Snapchat have to be addictive. The concept of network effects is nothing new, but it has scaled like never before with social platforms – Facebook being the prime example.
Network effect occurs when something becomes more attractive because more people are using it. To step outside of the social media realm for a minute, let’s consider video game consoles. If more gamers own an Xbox, it becomes more attractive for video game developers to develop games for Xbox. The more games that are developed for Xbox, the more attractive the Xbox console is to gamers. The cycle continues until Xbox becomes ubiquitous across the gaming industry.
With social media, it is even simpler. We want to be on Facebook because our friends are on Facebook. The more of our friends that are on Facebook, the more valuable Facebook is to us.
Network effects are exactly why social media companies have no choice but to run tests constantly to determine how to make their platforms as addictive as possible. In the attention economy, if they loosen the reigns on our attention for one minute, another tech company will swoop in and gobble up our time. If Twitter, for example, decided it existed for the good of society and no longer wanted to be addictive, the time we spend on Twitter would shift to other social networks that would optimize their way into our lives. Our friends would gravitate to a new platform, we’d go there because our friends are there, and the rest would be history. Every minute we aren’t on Twitter could be spent on Youtube, Facebook, Netflix, or – worst of all – sleeping, working, or engaging with people in the real world.