What does it mean that optimization has diminishing returns?

Each improvement gets harder and smaller — know when further optimization is not worth the cost.

Example: Vilfredo Pareto observed in 1896 that 80% of Italy's land was owned by 20% of the population. This pattern — later named the Pareto Principle — reappears everywhere: Microsoft found that fixing the top 20% of reported bugs eliminated 80% of crashes. The first 20% of optimization effort captures the vast majority of available gains. The remaining 80% of effort fights over scraps. A developer who spends two hours refactoring a critical hot path may cut response time by 40%. The next two hours might yield 5%. The next two, 0.3%. The function being optimized has not changed — but each unit of input now buys radically less output.

Try this: Select a system, habit, or process you have been actively trying to improve. Draw a simple chart: X-axis is total effort invested (hours, iterations, dollars), Y-axis is total improvement gained. Plot your best estimates for each round of optimization. Identify the inflection point — the moment where the curve bent from steep to flat. Now answer honestly: how much effort did you invest after that inflection? Calculate the cost-per-unit-of-improvement for your first round versus your most recent round. If the ratio exceeds 10:1, you are deep in diminishing returns. Write a one-sentence stopping rule you can apply next time you optimize anything.