Promptmkr

GraderExamples › Email

"Write a marketing email" — the prompt that returns filler, rewritten

Why a one-line marketing email prompt produces generic copy, and a reusable version with audience, offer, constraints and a ranking rule.

The prompt, as typed

Write a marketing email for our sale. Make it professional and persuasive.

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

18/100 · grade F

Clarity of context2
Structure3
Specificity2
Reusability0
Decision logic0

Rewritten

90/100 · grade A-

Clarity of context 5 +3
Structure 6 +3
Specificity 6 +4
Reusability 6 +6
Decision logic 5 +5

The rewritten prompt

Write a promotional email.

Audience: {{audience}}
Offer: {{offer}}
Deadline: {{deadline}}
Single desired action: {{cta}}

Output:
1. Three subject lines, max 45 characters each
2. Preview text, max 90 characters
3. Body of 120-160 words, one CTA, no sub-offers

Constraints: no exclamation marks, no "don't miss out", no countdown language
beyond stating the deadline once.

Rank the three subject lines by how clearly they state the offer, and give a
one-line reason for the top pick.

Why the rewrite works

The original asked for "an email" and got one — but nothing in it was decided by you. The rewrite fixes the deliverable count, the character limits and the banned phrases, so two runs a month apart are comparable instead of merely both being emails.

Common questions

Why does this prompt score badly?
It scores 18 out of 100 (F) because it leaves the model to decide 4 things you already know: what is on sale, for whom, and for how long; the single action the email should drive; word count and whether there is a subject line; what the brand never says.
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