Invariable Reinforcement

Our office manager Sophie passed me the phone. “It’s someone from Google,” she said. I raised an eyebrow. Perhaps this was an invitation to an event, or another chance to test prototype hardware, or something even more magical.

I unmute the phone. “Hello?”

“Hi, I’m Tim, from Google Digital Development. I’d love to talk about how we can help you promote your apps on the Google Play Store better.”

How disappointing — they were just selling Google search ads. I quickly made my excuses and hung up.

Three months later: “Hi Adrian! My name is Mike, I’m from Google Digital Development -”

Seven months: “Hey Adrian! I’m from Google Digital -”

Twelve months: “I’m Sean, I’m from Google Digi -”

To this day, it keeps happening and I keep getting my hopes up, like a child. Why don’t I learn that ‘Google’ on the phone equals ‘Irish guy cold-calling with ad sales’?

Because I haven’t told you about the times Google contacts us about actual interesting projects. It’s usually by email, but sometimes they do call. Not on a regular schedule, of course — but at random, unpredictable times.

This pattern of frustration mixed with intermittent success is essentially a variable reinforcement schedule. If you’ve read any article about addiction in the last twenty years, you’ll know that a variable reinforcement schedule can be used to make rats compulsively press a lever in the hope of getting another pellet of food; and that the same schedule could explain how addictive behaviour develops in humans.

Some people in the tech community act as if variable reinforcement schedules were occult knowledge, magic words capable of enchanting muggles into loosening their wallets. If only we could learn the secrets of variable reinforcement schedules, we could make them addicted to our new app — all those microtransactions, all those ad views, oh my!

So when people learn that I studied experimental psychology and neuroscience at Cambridge and Oxford — and that I run a company that designs health and fitness games — they are taken aback. They are fascinated. And then… they are disappointed, but only after I tell them that the principles of variable reinforcement schedules and operant conditioning can be learned by a dedicated student in a few hours. Moreover, if experimental psychologists were all capable of making the next Candy Crush, they wouldn’t spend most of their time complaining about the quality of tea in the staff common room.

That doesn’t mean that variable reinforcement schedules are bunk, though.

Variable reinforcement schedules help explain why I spend an hour a day mindlessly checking Gmail, Metafilter, Reddit, Twitter, and Hacker News. Even when I know, with 99% certainty, that nothing interesting will have happened in the 15 minutes since I last checked them, I still type Command-R — because maybe this time I’ll get lucky.

More broadly, it’s why we pay attention to the constant interruptions that plague our screens — there’s no cost to the person sending the interruption, and occasionally, it’s of real interest to us.

This plague has its origins from the dawn of email, but this year’s it’s broken out into the mass consciousness, at least if you measure by rapidly proliferating NYT opinion pieces and TEDx talks. It’s most recently been discussed by Tristan Harris (here’s his TEDx talk); Harris is a design philospher at Google, but he originally arrived there after they acquired his company, Apture, back in 2011. His particular interest right now is the Time Well Spent movement.

The purpose of the movement is to encourage the design of products and tools that allow users to make informed choices about how they spend their time. In other words, a user visiting a ‘good’ YouTube might be asked how long they want to watch videos for. After their time is up, the website would tell them to do something more useful and come back later.

I’m sympathetic to Time Well Spent, not least because their success would save me a lot of time. But on balance, I’m skeptical that companies can be convinced to engineer their products to make them less compulsive out of the goodness of their hearts, any more than advertisers and publishers can be convinced to reduce the number of obnoxious and unsafe ads out there.

I’m happy to be proven wrong, but let’s put it this way: Harris works at Google, and I don’t see any friendly ‘how long do you want to spend surfing the web?’ dialogs in Chrome. No, perhaps we should take matters into our own hands — like we did with third party ad blockers.

While it took ad blockers many years to gain traction, they’re now used by a significant percentage of browsers — at least 15% in the US, 20% in the UK, and 25% in Germany. The advent of Content Blocking in iOS may see those numbers continue to grow. So it’s tempting to think that a similar strategy, centred around browser extensions, could help disrupt the many variable reinforcement schedules that bind our attention.

In fact, many such apps and extensions exist, like Freedom, StayFocusd, and LeechBlock. Let’s call them ‘compulsion blockers’. Not all compulsion blockers are apps — at university, my friend Alex’s version of a compulsion blocker was giving me his network cable while he was trying to write an essay.

Compulsion blocker apps have not made much of an impact. You’d know if they had, because the wailing from app developers and games companies would be deafening. It’d make publishers’ complaints about ad blockers seem like a kitten’s meow — just imagine if 20% of people used compulsion blockers to reduce their Facebook or Tumblr or YouTube time. It’d be the bonfire of the unicorns!

Why haven’t they been more successful?

1. Many people actually enjoy browsing Facebook and YouTube, thank you very much. And how dare you say that they’re wasting their time refreshing Reddit every five minutes!
2. While some people (e.g. the readers and author of this article) may believe that compulsive browsing on computers is the main problem, the truth is that compulsive smartphone usage is much worse. And making compulsion blockers for smartphones is really, really tricky.

It’s technically possible to a create compulsion blocker for Android phones; some kind of custom launcher app that replaces the home screen and can monitor and block the usage of any app or website (just imagine the permissions list you’d need!) Unfortunately, custom home screens aren’t very popular beyond power users. Even the full might of Facebook wasn’t enough to make their custom Home launcher a success. People just don’t seem to care that much.

But it gets worse: it is literally impossible to make a compulsion blocker for the iPhone and iPad. Third-party developers simply cannot make apps that block or control the behaviour of other apps, and any attempts to make an end-run around Apple’s locked-down App Store distribution model have not been successful. I can’t imagine this will change any time soon, either.

If a technological solution can’t be found on smartphones, perhaps we need to go further up the stack. Maybe when augmented reality glasses finally arrive, we can use them to blank out our phones whenever we try to open up Candy Crush for the twentieth time!

But our technological masters — Apple, Facebook, Google, Microsoft — they aren’t dummies. They realise that augmented reality and virtual reality represent the ‘final compute platform’ that could subsume all other computing and display devices. They would do anything to control and monetise that future, including prohibiting developers from making apps that control other apps, just like Apple does. It’ll be the war to end all platform wars.

Let’s summarise: compulsion blockers aren’t popular on desktops, they’re neglected or prohibited on smartphones, and the same may be true on future platforms as well. All hope is lost.

Or is it?!!!

There are other things in this world that are highly addictive. They’re called drugs. We even have ‘drug blockers’ like naltrexone, which block the action of opioids on a molecular level. The slow-release injectable version of naltrexone is called Vivitrol, and can be used to control heavy opiate and alcohol addictions.

Naltrexone and Vivitrol aren’t household names because most people aren’t dangerously addicted to drugs or alcohol. They aren’t much used as a preventative measure either, because a lot of people enjoy taking drugs and drinking alcohol, thank you very much.

Likewise, most people aren’t dangerously addicted to Facebook, so they don’t feel they need a compulsion blocker. For my own part, I don’t use a one because my behaviour doesn’t seem too bad, and I also quite enjoy browsing the web.

Let’s assume that it gets worse, though. Not a foolish assumption given that there are thousands of people spending billions of dollars, trying to make us compulsively use their apps and websites. Maybe the hour a day I spend checking websites goes up to two or three hours a day, in which case I will be highly motivating to get myself a compulsion blocker.

Unfortunately, compulsive experiences generate a lot of cash. The people behind those experience will therefore be highly motivated to circumvent any blockers — consider the phenomenon of advertisers paying popular ad blockers to let their ‘acceptable ads’ through. Yes, there is no escaping capitalism.

For that reason, if we want to genuinely reduce compulsive behaviour, we can’t simply ask VC-backed or publicly-owned companies to play nice. We can’t even ask their employees to play nice; there are just too many smart people out there who are more than happy to take Facebook or Google or Supercell’s $250,000 salaries a year and turn a blind eye to questionable design practices.

Here’s what we can do: we can outcompete them. There’s a reason why we don’t spend literally all of our time on computers or smartphones messing about on Facebook or Candy Crush, and that is because there are better things to do. It might be reading Station Eleven, or watching Mad Max: Fury Road, or playing Life is Strange.

We also need tools and devices and venues that allow us to experience these things without interruptions. Lately I’ve made a habit of going to the cinema to watch movies — it helps me focus on the movie rather than checking my phone, and I come out appreciating it more. Likewise, I bought a Kindle Paperwhite so I can more clearly delinate my time between browsing the web and reading a proper book.

You can make money with some of these things. Not unicorn money, perhaps, but certainly a lot. More importantly, a good book, a good movie, a good game — these things are all worth of creation and consumption in and of themselves.

A good movie or book doesn’t compel us with a variably reinforced schedule to visit it again and again and again, until we’re exhausted. No, they compels us to come back because they’re well-made, right from their beginning to their very satisfying, and very final, end.