On marginal costs

Marginal cost is the expense in building one additional unit. In theory, as you learn how to do something, this cost becomes lower. Often this is called economies of scale or experience curves.

My assertion is that the theory doesn’t match up to reality in (maybe) all fields that don’t involve producing physical things in a factory or a farm.

Starting with dinner. I love cooking. In theory, the marginal effort in me cooking for just myself and for my wife of nearly a decade is nearly zero. In practice, as someone who cooks ~5 meals a week, I’ve cooked just for myself maybe 5 times total. The marginal effort put into producing a singular unit always comes with significant capex.

Whenever we have friends over (~1/month), things are always special. Side dishes, dips, desserts, drinks. The marginal effort for more than 2 people (and their furry friends) jumps from half an hour to an entire afternoon. The experience curve spikes upwards at critical steps, it doesn’t monotonically keep going down.

Let’s pick another example, say acquiring customers by running instagram ads for an e-commerce firm. Each new campaign comes with a setup cost and learning phase which behaves like capex. Every time the ad creative reaches relative saturation for a demographic or for a number of impressions, acquiring more users requires more marginal spend, not less.

You’re still skeptical as one should be. Let’s look at progressive overload techniques to build muscle. Starting a new exercise has significant injury risk and hence, capex like cost. Once you cross that, marginal effort in each new rep is low till you come close to failure. The failure zone is ridiculous effort, causes near blackouts but it’s where gains happen.

One last one, I promise. Assuming the same % of mislabelled data, as ecommerce firms keep adding things to the catalog, search retrieval quality becomes worse unless some conscientious soul decides to actively work on the problem.

Ok, I break the promise, one last last one. Hiring employee number 1 in a team is a massive capex investment in rebuilding processes. As you hire more people, a communication tax is applied which slows everyone down. This tax rate increases with each new sub-unit formed and approvals needed.

One white flag for intellectual honesty. I’ve used the term ‘marginal cost’ and ‘capex’ loosely to describe the activation energy required for the next phase of growth, not just one identical unit. There are dis-economies to success and the premise needs to be challenged more often than it is.

Go forth and be depressed gentle one… you’ve reached here and earned it.