Variable Rewards

This part of psychology is all about how to get a user to keep coming back to an application or service, to have the urge to use it – what we might otherwise call addiction.

Variable rewarding is a brain hijacking technique designed to create compulsion loops – habits that give you the urge to go back and check for ‘a reward’ (maybe a like or a comment). This technique has long been used by slot machines which are known to be highly addictive (they are known as ‘pokies’ in Australia). Variable rewarding is built into many applications and services – the aim is to make you feel compelled to use them as often as possible. This is sometimes called “The Hook Model”.

Explain Like I’m Five

Your brain is hardwired to try and predict the outcome (and reward) of an action that you take. When you are expecting a reward, you often feel anticipation and excitement. If you receive a reward that is unexpected, that excitement level is even greater.

In terms of technology, the action is using the product (that might be checking your phone, checking Facebook or Instagram) and the reward is some form of positive affirmation, like a notification, a like or a comment. You say “checking Facebook” because your brain wants to know whether or not there is a reward for you. The uncertainty here is key – you are curious to know what has happened.

These services can randomise the reward – every time you visit it is different. If you refresh the page again (like pulling the slot lever) the content changes. Sometimes it shows you notifications, likes and comments. This not knowing keeps you coming back, forming a habit to check your phone or app regularly to see if you have a reward. This behaviour is linked to a chemical called dopamine, which is the same one impacted by cocaine and alcohol (deep dive below).

The use of The Hook Model is widespread. Books have been written about it (such as Hooked: How to Build Habit-Forming Products by Nir Eyal) and there are companies that allow you to use AI to deliver random rewards to your users. One such company is Boundless Mind, which shrewdly changed its name from Dopamine Labs.

Habits are not inherently bad or negative. You can use variable rewards to build positive habits like studying or exercising. It is important to be aware when products are deploying this method against you and hijacking your mind. Maybe it is for what they consider a noble reason, like helping you to form a meditation habit. But that becomes ethically tricky because someone has to decide what behaviour is ‘noble’ or desirable, and trickier still when these methods are deployed in products used by children.

You should be aware of this method and try and consciously think about why you are unlocking your phone or opening apps. Are you hooked?

Deep Dive

This particular area of psychology is known as ‘operant conditioning’ and a lot of what we know can be traced back to B.F. Skinner and his experiments with ‘Skinner’s Box’. He came up with some ideas around ‘reinforcement’, where behaviour that is reinforced is strengthened and repeated, while behaviour that is not reinforced slowly becomes extinct.

Skinner set up an experiment to test different types of reinforcement. He put a rat in a box with a lever. When the rat pressed the lever, it delivered a pellet of food (continuous reinforcement). He tried quite a few combinations of positive reinforcement, negative reinforcement and punishment. He measured how many times the rat pressed the lever and how long before it gave up.

He found that ‘variable reward’ was the strongest type of reinforcement – this is when the rat pressed the lever and sometimes it received a pellet of food and sometimes it didn’t. This behaviour with variable reward didn’t become extinct – essentially it became compulsive. You can watch Skinner himself describing this here.

So why does this operant conditioning work on humans, how does it work physiologically? This brings us into the realm of neuroscience and the work of dopamine within the reward pathways of our brain. Read about it in more depth here.