Late in the summer of 2013, my friend Brad and I were at an ice cream parlor and decided to get milkshakes. There were two options: a 12oz^[0] milkshake for $3, or a 24oz milkshake for $5. Brad opted for the smaller one, but I got the larger one. After all, it was clear that the larger one offered more value per dollar.

We were walking down the street, eagerly drinking our milkshakes. Brad stopped by a trashcan. He held his small milkshake — still half-full — and threw it out. “Why’d you throw it out? Was it no good?” I asked.

“No, it was great.” responded Brad. “But when I buy a milkshake, I really just want the first sip.” Of course he was right. All the flavor and pleasure lies in the first few sips. After that, you’re just absent-mindedly slurping your shake and freezing your hands. An economist might say that a milkshake has rapidly diminishing marginal utility.

Brad had optimized happiness for money. He knew that the first sip would be delicious, and the rest wouldn’t really do anything. To him, the big shake had nothing to offer over the small shake. I, on the other hand, did the classic value-for-money optimization. I figured out which deal got me more milkshake per dollar, and took that one. Even though my ultimate concern was which deal made me happier, I didn’t give that any thought. The way I made my choice was typical of a trader: buying as much as possible as cheaply as possible, as if I’m going to sell it later on. By contrast, if I had been a smart consumer, I would have recognized that all I wanted or could use were the first few spoonfuls of milkshake, and bought a small one.

In Thinking, Fast and Slow, Daniel Kahneman discusses what he calls attribute substitution. The gist is that when we’re confronted with a difficult question, we often subconsciously replace it with a similar, but easier question, especially if we’re under pressure. In this context, optimizing for the amount of milkshake per dollar is an easy question. It’s just arithmetic. But when you optimize for happiness, you have to think quite seriously about your own preferences: what makes you happy, quantifying that happiness, and finally weighing it against something else, like money or time. I think that’s really hard. So it’s no surprise that I swapped the hard question for the easy one, and thus optimized for the wrong variable.

This raises a question: do I optimize for the wrong variable in other contexts, too? I’m convinced I do it all the time. When choosing between options, I ultimately have to think about questions at the most fundamental level: will this make me happy? Is this something I want to do with my life? Do I know myself well enough to make this assessment? Do I truly like what I think I like?^[1] Regrettably, such questions are almost prohibitively difficult. They’re hard to engage, even when not under pressure. In turn, it’s hard to carry out the appropriate optimization, and easy to instead optimize for something else — which might feel similar, but may actually be totally unrelated.

In the spring of 2015, I was preparing to graduate from college. I met with a friend/mentor, and pitched to him my career plan. He listened carefully, and said: “I think you’ve spent too much time solving the problem, and not enough time choosing the problem.” He was right. The question of paramount importance is, out of an infinite set of potential goals, which ones to pick. That’s hard. As such, I had basically ignored it, picked goals without much thought, and instead thoroughly solved the much easier question of how to accomplish those goals. I had wasted my time.

It’s a common disappointment to find an achievement unrewarding or meaningless after spending a great deal of time and effort working toward it. Usually, it means you chose the wrong goal,^[2] i.e. optimized for the wrong variable. I think the fundamental cause of this is attribute substitution: instead of answering the essential question about self-knowledge, to answer a substitute that fails to get to the heart of the matter.

[0] About 350mL.

[1] I think it’s important to carefully examine the things that you think you like or bring you happiness. David Foster Wallace suggested that we think we like certain things only due to effective advertisement or some other form of dependency, which ultimately distracts us from the things that actually make us happy.

[2] Setting good goals generally seems to be a tricky business. Scott Adams suggests in his book How to Fail at Almost Everything and Still Win Big that a better approach is by following systems that will, in the long run, naturally move you toward success. For example: “losing weight” is a goal that requires a lot of willpower and can seem pretty arduous. “Eating healthily” is a system that is easier to settle on, and it implies a plethora of good effects — not just losing weight. “Become a great philosopher” is again a lofty goal that will more likely than not lead to disappointment, but “learn about philosophy every week” is a system. Goals are tough because until we’ve achieved our goals, we feel as if we’re falling short of them, and the satisfaction from the moment of victory usually doesn’t last long. Systems, on the other hand, can continuously reward us. I agree with Adams: setting up flexible, but well-directed systems is more likely to produce long-term success than the staccato climb from goal to goal.