Second-order effects
The first effect of a decision is the one you intended. The second is how people and systems react to it. A discount brings in customers — and also teaches them to wait for the next discount.
The idea
First-order thinking stops at the immediate result. Second-order thinking keeps going: once this change is in place, what will everyone affected do differently? Those reactions can reinforce the original goal, cancel part of it, or create something nobody planned.
The reactions matter because systems are full of people and processes that adapt. Prices change behaviour, convenience changes habits, and spare capacity tends to get used.
The coal question
In 1865 the economist William Stanley Jevons studied British coal use and noticed something counter-intuitive. James Watt’s improvements had made steam engines far more efficient, so each engine needed much less coal. Yet total coal consumption had grown sharply. Cheaper steam power had become worthwhile in mines, mills, railways and ships where it had not paid before. The efficiency gain was real; the economy simply found many more jobs for it.
The general pattern is now called the rebound effect. When something becomes cheaper to use, people tend to use more of it, which takes back part of the expected saving. LED lighting is a modern example: it needs far less electricity for the same light, and cheap light has encouraged more lighting — bigger displays, more decorative and outdoor lights.
How large the rebound is depends on the technology and the market, and economists still debate the size for many cases. In most everyday situations it takes back only part of the saving. Jevons’ coal was an extreme case, where it took back more than all of it.
Second-order effects are not always bad. Making a slow automated test suite faster has an obvious first-order benefit: less waiting. The second-order benefit is often larger. When tests are fast, developers run them far more often, so problems are caught earlier and fixed while they are still small.
How to use it
- Write down the immediate, intended result.
- Ask “and then what?” at least twice.
- At each step, name who changes behaviour — users, suppliers, competitors, colleagues, yourself — and over what time scale: days, months or years.
- Look for loops. Does the reaction feed back into the original cause, making it stronger or weaker?
- Pick one number that would show the second-order effect early, and watch it.
Where it misleads
Chains of consequences become speculative quickly. Two steps are usually enough; by the fourth step you are mostly writing fiction.
It can also become an excuse for doing nothing, since every action has some imaginable downside. Use it to anticipate and watch the reactions, not to veto every change. And effects have sizes: a second-order effect that is real but small should not outweigh a large first-order one.
In one line: every change invites a reaction — work out the reaction before you count the benefit.