Stepwise Decomposition

Breaks down a problem into ordered intermediate steps that can be solved before the final answer, to expose reasoning and prevent premature conclusions.

Also known asChain-of-Thought PromptingStep-by-Step ReasoningSequential Reasoning

Problem

One-shot answers hide intermediate reasoning, making errors hard to detect and encouraging premature conclusions.

Solution

Require ordered intermediate steps before final commitment. Externalized steps function as visible working memory and local checkpoints, improving inspectability and robustness.

Example

One worked example. The book has 2–4 per technique.

Prompt

We have 500 active subscriptions. 300 are on annual plans at $948/year and 200 are on monthly plans at $99/month. What is our monthly recurring revenue? Solve step by step.

Response

Step 1: Monthly plan revenue. 200 subscribers × $99/month = $19,800/month.
Step 2: Annual plan revenue, normalized to monthly. $948/year ÷ 12 = $79/month per subscriber. 300 × $79 = $23,700/month.
Step 3: Total MRR. $19,800 + $23,700 = $43,500/month.

Final answer: Monthly recurring revenue is $43,500.

Techniques

Concrete ways to implement Stepwise Decomposition. Each technique fits a different situation.

01
Step-by-Step Instruction
Tell the model to work through ordered steps before giving a final answer, making reasoning visible with minimal prompt overhead.
02
Demonstrated Reasoning
Provide worked examples that show how the task should be decomposed, then present the new problem for the model to solve in the same style.
03
Incremental Stepping
Request one step at a time across multiple turns so the user or system can inspect and redirect between steps.

Full treatment in the book