Which Prompt Is Better? A Decision Checklist
Seven questions that decide between two prompts — audience, format, length control, constraints, criteria, ambiguity, and contradictions.
Overview
"Which prompt is better" has a checkable answer most of the time. Better prompts define who the output is for, what shape it takes, how long it should be, what to avoid, and how to tell when it's right. Weaker prompts replace those decisions with adjectives like "detailed" and "high quality". This checklist turns that into seven concrete questions, and the loaded example shows a pair where the checklist makes the winner obvious in one pass.
Workflow
-
Run the checklist questions
Audience? Format? Length control? Constraints? Success criteria? Vague wording? Contradictions? The comparator checks all seven automatically.
-
Compare the loaded pair
The example shows a 'make it good' email prompt against one that answers every checklist question. Watch where the scores split.
-
Check the close-call case
If scores land within a few points, the verdict says so — then the category table is your tiebreaker, not the overall number.
-
Keep the checklist habits
The improvement suggestions are the checklist in action: each one is a missing answer to one of the seven questions.
Why This Works
- Concrete questions beat taste: two people running the same checklist reach the same verdict
- Adjectives like 'good' and 'detailed' fail the checklist because the model can't act on them — the score reflects that directly
- A checklist scales: the same seven questions work for emails, code review prompts, and research prompts alike
Best for
- Decisions that keep recurring because nobody can articulate why one prompt feels better
- Teams that want a shared, repeatable definition of prompt quality
- Quick pre-flight checks before a prompt goes into an automated workflow
Not for
- Deep prompt rewriting — the checklist decides, it doesn't redraft
- Cases where both prompts target different tasks — compare like with like
Use cases
- Reviewing a teammate's prompt against your own before standardising one
- Teaching a team what separates a strong prompt from a weak one, with scores as evidence
- Auditing a prompt library two entries at a time to keep only the stronger variant