Promptmkr

GraderExamples › Content

"Write a case study" — the prompt that produces a testimonial, not evidence

Why an open case-study prompt returns vague praise instead of a result, and the rewritten version built around a measurable before and after.

The prompt, as typed

Write a case study about one of our customers. Make it compelling and professional.

What it leaves the model to guess

Graded before and after

Both versions below were scored by the same rubric the free prompt grader runs.

As typed

8/100 · grade F

Clarity of context0
Structure0
Specificity2
Reusability0
Decision logic0

Rewritten

90/100 · grade A-

Clarity of context 8 +8
Structure 10 +10
Specificity 8 +6
Reusability 6 +6
Decision logic 5 +5

The rewritten prompt

Write a customer case study for {{audience}} at the {{funnel_stage}} stage.

Customer: {{customer}} — {{customer_context}}
Situation before: {{before_state}}
What changed: {{intervention}}
Measured result: {{measured_result}}
Objection this case study must answer: {{main_objection}}

Tone: {{tone}}. Channel: {{channel}}.

Output, 550-700 words:
- Headline stating the result, max 12 words
- Situation section: the before state in the customer's terms
- What changed: what was actually done, in order
- Result: the measured outcome, with the number stated once and not repeated
- One pull quote of no more than 25 words

Constraints: no claim that is not supported by {{measured_result}}. No
"partnered with", "seamless", "excited to announce". Do not describe the
result as "significant" — state it.

Write 3 headline options and rank them by how directly the result answers
{{main_objection}}, with one line explaining the top pick.

The case study succeeds if a sceptical reader can say what changed, by how
much, and over what period. Check the draft against that test before finishing.

Why the rewrite works

Most case-study prompts produce praise because nothing in them requires evidence. Naming the before state, the intervention and a measured result turns the model's job from describing a relationship into reporting a change — and the final check gives it a test it can actually apply.

Common questions

Why does this prompt score badly?
It scores 8 out of 100 (F) because it leaves the model to decide 4 things you already know: the measurable before and after; which objection the case study is meant to answer; who reads it and at what stage; length, structure and what cannot be claimed.
Does a longer prompt always score higher?
No. The rubric rewards named inputs, a fixed output format and an explicit rule for choosing between options. Extra prose without those adds length and no score.
Will this rewritten prompt work in ChatGPT and Claude?
Yes. It is plain text with placeholders you fill in before running, so nothing in it is specific to one model or vendor.

More worked examples

More bad prompt examples · How to write better AI prompts