Reinforcing Habits

Epistemic status: hypothesis based on >10 anecdotal examples

Every month or so, I’ll get a client asking why their attempts to start a habit failed. They want to have an automatic action requiring minimal willpower. The client is usually familiar with at least one habit-building model. Most commonly Charles Duhig’s Cue, Routine, Reward loop (in “The Power of Habit”) or CFAR’s Trigger-Action Plans (TAPs).

Their model may go like Duhig’s story: he wanted to change his habit of eating a cookie each afternoon (motivated by watching the scale creep up). So he identified his cue (the time of day around 3pm), planned a new routine to replace the old one (talk to colleagues for ten minutes), and had a new habit. My client usually wants to know what thwarts their attempts to do likewise.

What’s missing from that example?

I read The Power of Habit many years ago, and honestly didn’t remember Duhig’s cookie story any better than my clients often do. So I was surprised when I revisited it and found Duhig actually did a series of experiments to find what the reward was. Something sweet? Nope, eating a candy bar at his desk didn’t feel great. Just taking a break? Nope, taking a walk outside didn’t cut it. Talking with friends? Yeah, that felt rewarding. Based on his experiments, we can guess that getting a cookie each afternoon was a means for getting him to talk to his colleagues while eating it. Because the reward here was socializing, he could build a new habit that didn’t use the cookie as an intermediate step.

This reward step is often neglected by my habit-struggling clients. They want the low-effort, automatic aspects of habits. They lack, however, anything to make those behaviors sticky.

I think it might help to reframe habits as repeatedly-reinforced behaviors. Our brains, often subconsciously, have tied a particular action to some cue after repeatedly having that action rewarded. Simple patterns of cue, action, and reward in close proximity get reinforced, such as “it’s mid afternoon → I’ll get a cookie → rewarded by social connection.”

Intentionally designing good habits is hard. By default, our unconscious habits are selected for rewarding behaviors. For example, when I’m feeling blue, eating chocolate and hugs make me happier. I don’t need to train myself to eat chocolate when I feel blue—this happens quite easily!

On the other hand, you have to intentionally find the reward when you’re trying to kickstart a habit. In particular, you need to find the reward if you want your habit to happen automatically without willpower each time. A random desired behavior may not have an immediate reward, so you need to experiment. (I’m not guaranteeing that all behaviors can be made into habits—there’s a reason that doctors recommend “whichever exercise you’ll actually do” rather than a specific fitness-optimized routine.)

Two similar actions can cause very different experiences due to small differences specific to you. Pullup hangs and RSI wrist stretches were both small actions that I repeated for brief reps three times a day. The pull-up hangs were motivating because I could see myself improve day to day—an extra second here, three seconds longer there. The RSI stretches quickly became demotivating because I couldn’t see myself making any progress even after a couple weeks of consistent use.

Similarly, 7am yoga required financial penalties to get me into downward dog before the world warmed up. Walking a mile in the peaceful evening while thinking or calling a friend was a piece of cake in comparison.

The best rewards are natural consequences of the action—i.e. the experience of doing the action reinforces the behavior. The reward might be enjoyment of the action, seeing progress toward a goal, a social status boost, consistency with your sense of self, connection, release from a worry, etc. Note, all of these are gut-level feelings, not “shoulds”. A System 2-level sense of “I should…” doesn’t seem to have the same rewarding effect. You actually have to find what feels rewarding. If you can identify and increase the reward, you can make the habit easier to sustain. This implies that the best way to deliberately change/start habits is to choose new habits with immediate positive outcomes, and make those benefits salient.

On the other hand, you can also try to tack on a reward that doesn’t inherently come with the action, such as fist pumping the air or using financial penalties. Arbitrary rewards can be quite useful (particularly in situations when you just have to push through something unpleasant). However, financial penalties and other arbitrary rewards fall apart if you stop applying a bit of willpower to set them up each time. Rewards with a self-coercive element are also more draining/stressful to use for many people. (Not surprising—the rewarding habit is more pleasant than a financial penalty.)

One popular example is temptation bundling. The idea is you only give into a temptation while also doing a desired activity, such as only watching TV while exercising. However, a common outcome is someone tries to watch TV exclusively while exercising, only to have some part of their brain point out that nothing is stopping them from watching TV in bed…

In contrast, here are few examples of people using natural rewards to reinforce habits:

According to Nate Soares, “When I was quite young, one of the guests at our house refused to eat processed food. I remember that I offered her some fritos and she refused. I was fairly astonished, and young enough to be socially inept. I asked, incredulous, how someone could not like fritos. To my surprise, she didn’t brush me off or feed me banal lines about how different people have different tastes. She gave me the answer of someone who had recently stopped liking fritos through an act of will. Her answer went something like this: ‘Just start noticing how greasy they are, and how the grease gets all over your fingers and coats the inside of the bag. Notice that you don’t want to eat things soaked in that much grease. Become repulsed by it, and then you won’t like them either.’”
Tara Mac Aulay: “I found that going to lift weights with friends is surprisingly good, because I get to have a good chat with them for an hour, and it’s not strenuous enough that you can’t have a conversation. But if I was to go on a bike ride with friends or something else where you can’t talk, it’s not as nice. And my main exercise is probably just walking and dancing. I go out dancing a lot on my own, to go and see music artists that I enjoy, and I just dance like a crazy person until I’m really tired and then I go home, and that’s amazing.”
I struggled for a while to brush my teeth consistently. I eventually paid mindful attention to the feeling of stuff on my teeth. The little layer of course film when I hadn’t brushed after eating, and the polished silk of freshly brushed teeth. I started getting annoyed at the texture on my teeth, and then brushing was easy.

Experimenting and paying close attention to what feels rewarding seem to be common elements behind the successes above.

A couple ideas if you want to try using natural rewards to build habits for yourself:

Run experiments to find what you enjoy enough to easily make a habit. You can track these formally, or just note which are easier to do repeatedly.
Use Soares’ technique of focusing on the experience of minute details that attract you to habits you want, or make unwanted habits less desirable. For example, pay attention to how much better you feel when you don’t have an important email hanging over your head, compared to when you were procrastinating.
Use this CBT Pleasure Predicting Worksheetto increase your awareness of how much you enjoy activities. This sheet works by highlighting discrepancies between expected and experienced enjoyment.