To think at the margin is to think ahead. Here we use “marginal” to mean “additional”. The utility derived from one unit of a thing rarely perfectly doubles with two of that same thing. To be more specific, imagine the pleasure experienced in the first bite of a rich chocolate cake. The next will provide slightly less satisfaction, and by the time you have finished the cake, the idea of another bite may seem unappealing. When you’re thinking at the margin, you’re considering what an additional action means for you.

The marginal product is that which we obtain from extra investment. Let’s stipulate that another hour following up with a client will motivate them to order 12 more widgets from your company. The marginal product of your labour here is 12 because an extra hour sold this many more widgets.

The marginal cost of selling an extra widget was 5 minutes (1/12th of an hour) of your labour. Each sale of one costs you 5 minutes. After selling 12, this cost might increase to 10 minutes, meaning you may only sell 6 more after another hour of labour.

Another example; a t-shirt factory can produce 1000 t-shirts an hour at max capacity. While manufacturing 800 t-shirts, the marginal cost of producing 200 more is whatever the materials cost because the capacity is there. Making 300 more t-shirts would require a new factory, employees to work in it, and materials. The extra 100 above capacity may not justify that cost.

Thinking at the margin is a solution to the sunk cost fallacy. Economists sometimes say “sunk costs are sunk”. After having invested time, effort, and energy into an endeavour, it’s difficult to accept that the best next move is to pursue a different course of action. Waiting for a bus to avoid walking 45 minutes through the rain makes sense for most people. However, 30 minutes into waiting for a bus, one might wonder whether or not giving up and walking is the best next move. Somebody thinking at the margin would assess how likely the bus will turn up in the next 30 minutes. If the bus is 80% likely not to arrive, the economist will embrace the discomfort and start walking. The invested 30 minutes have already passed; there is no use compounding the error. Think about what is next, not what has been.