Email Subject Line Prompt Templates for ChatGPT: A Tested Reference Library

Subject lines are one of the highest-leverage places to use ChatGPT in email marketing — the output is short, easy to evaluate, and you can generate 20 variants in under a minute. But the quality gap between a well-constructed prompt and a vague one is significant. Generic prompts produce generic subject lines. This reference library covers prompt templates that have been tested against specific campaign types, with notes on what to adjust and where the model tends to go wrong.

What Makes a Subject Line Prompt Actually Work

ChatGPT doesn't know your list, your brand voice, or what your last three campaigns looked like. The model fills gaps with defaults — and those defaults tend toward marketing clichés. Subject lines that say "Don't miss out" or "You won't believe this" aren't bad because ChatGPT is bad; they're bad because the prompt didn't constrain the output.

A functional subject line prompt needs four inputs at minimum:

• The campaign type (promotional, transactional, nurture, re-engagement, event)
• The specific offer or message — not just "sale" but "30% off winter coats, ends Sunday"
• The audience segment and what they already know or have done
• Tone constraints — whether to avoid urgency language, emoji, questions, or a specific style

Adding a character limit also helps. Without one, ChatGPT frequently outputs subject lines that are 70–90 characters — fine for desktop, but clipped on most mobile clients. A 50-character target forces the model to prioritize.

Template Library by Campaign Type

1. Promotional / Sale Email

Use this when the email's primary purpose is driving a click to a time-sensitive offer. The template forces specificity on the offer and deadline, which reduces the model's tendency to pad with vague urgency.

You are writing email subject lines for a [BRAND TYPE] brand.

Campaign: Promotional sale email
Offer: [SPECIFIC OFFER — e.g., "30% off all outerwear"]
Deadline: [DATE OR TIMEFRAME — e.g., "ends Sunday midnight"]
Audience: [SEGMENT — e.g., "past purchasers who haven't bought in 90 days"]
Tone: [e.g., "direct and confident, no exclamation marks, no emoji"]
Length: Under 50 characters

Write 10 subject line options. Do not use the phrases "don't miss," "last chance," or "hurry." Vary the angle: some should lead with the discount, some with the product, some with the deadline.

2. Re-engagement Email

Re-engagement subject lines have a narrow job: get an inactive subscriber to open without feeling like a guilt trip or a desperate plea. The model's default here is either overly casual ("We miss you!") or weirdly formal. The template below constrains both directions.

You are writing subject lines for a re-engagement email.

Context: The subscriber has not opened or clicked in [TIMEFRAME — e.g., "6 months"].
Brand: [BRAND NAME OR TYPE]
What's new or different: [e.g., "new product line launched since they last engaged"]
Tone: [e.g., "low-pressure, curious, not apologetic"]
Length: Under 45 characters

Write 8 subject line options. Avoid: "We miss you," "Come back," "It's been a while." Try angles like: what's new, what they might have missed, a direct question, or a soft incentive reference.

The "what's new or different" field is the most important variable here. If you leave it blank or vague, the model has nothing to anchor on and reverts to sentiment-based subject lines.

3. Nurture / Educational Email

Nurture emails don't have a hard offer, which makes them harder to write subject lines for. The temptation is to describe the email's content — "5 tips for better X" — but that pattern is overused and trains readers to deprioritize. This template pushes toward curiosity or specificity instead.

You are writing subject lines for a nurture email in a [INDUSTRY] marketing sequence.

Email topic: [SPECIFIC TOPIC — e.g., "how to reduce cart abandonment with email timing"]
Audience stage: [e.g., "aware of the problem, evaluating solutions, not yet ready to buy"]
Tone: [e.g., "practical, peer-to-peer, no hype"]
Length: Under 55 characters

Write 8 subject line options. Avoid numbered list formats ("5 ways to..."). Try angles like: a specific insight from the email, a question the reader is likely asking, a counterintuitive claim, or a concrete result.

4. Event / Webinar Invitation

You are writing subject lines for an event invitation email.

Event: [EVENT NAME AND FORMAT — e.g., "live Q&A webinar on B2B LinkedIn ads"]
Date/time: [DATE AND TIME]
Speaker or host: [NAME AND CREDENTIAL if relevant]
Audience: [e.g., "B2B marketers at companies with 50–500 employees"]
Tone: [e.g., "professional but not stiff, no exclamation marks"]
Length: Under 55 characters

Write 8 subject line options. Vary the angle: some should lead with the topic, some with the speaker, some with the format (live, interactive), some with the outcome the attendee will get.

5. Transactional / Post-Purchase

Post-purchase emails have high open rates by default, but the subject line still matters for setting tone and driving secondary actions (review requests, cross-sells, referrals). The template here focuses on making the subject line feel personal rather than automated.

You are writing subject lines for a post-purchase email.

Purchase context: [e.g., "customer just bought a premium coffee subscription"]
Email purpose: [e.g., "order confirmation with a referral offer"]
Brand tone: [e.g., "warm, conversational, no corporate language"]
Length: Under 50 characters

Write 6 subject line options. The subject line should feel like it's from a person, not a system. Avoid: "Your order has been placed," "Order #[number] confirmed." Try angles that acknowledge what they bought, set expectations, or hint at the referral offer.

A/B Testing Prompt: Generate Paired Variants

If you're running split tests, the most useful output isn't a list of 10 random variants — it's pairs that test a specific hypothesis. This prompt generates structured A/B pairs with a stated test rationale.

You are helping design A/B tests for email subject lines.

Campaign: [CAMPAIGN TYPE AND BRIEF DESCRIPTION]
Current best-performing subject line (if known): [OR "none"]
Test hypotheses to explore: [e.g., "question vs. statement," "benefit-led vs. curiosity-led," "with vs. without sender name"]

For each hypothesis, write one Version A and one Version B subject line. After each pair, write one sentence explaining what the test would tell you if Version A wins vs. Version B wins. Keep all subject lines under 55 characters.

Tone Calibration: Getting the Model to Match Your Brand Voice

Tone instructions like "professional" or "friendly" are too vague to constrain the model reliably. More useful approaches:

Pasting 3–5 real subject lines from your archive as examples is the most reliable method. The model picks up on structural patterns, word length preferences, and stylistic tendencies that are hard to describe in abstract terms.

Known Failure Modes

• Urgency default: Even when not instructed, the model gravitates toward urgency framing. Explicit "avoid" instructions are needed for most promotional prompts.
• Character count drift: The model interprets "under 50 characters" loosely. Some outputs will be 55–65 characters. Count manually or ask for a character count next to each option.
• Repetitive angle clustering: When asked for 10 variants, 3–4 will often use the same structural angle. Adding "vary the angle" and naming specific approaches reduces this, but doesn't eliminate it.
• Hallucinated specifics: If the prompt includes vague context (e.g., "a software product"), the model may invent specific product names, features, or stats in subject lines. Always provide the actual specifics.
• Over-personalization tokens: The model sometimes outputs subject lines with merge tags like {{first_name}} even when not asked. These need to be removed or validated against your ESP's actual tag format.

Adjusting These Templates for Different Contexts

The templates above are starting points. Three variables that consistently change output quality:

1. Industry specificity. A SaaS nurture email and a DTC retail email need different defaults. Adding "[INDUSTRY] audience" in the prompt shifts vocabulary and reference points.
2. List temperature. Cold outreach, warm leads, and loyal customers warrant different tones and different levels of assumed familiarity. Name the relationship in the prompt.
3. Platform constraints. If your ESP previews preheader text next to the subject line, note that in the prompt. The model can then avoid repeating the same information in both.

What These Prompts Don't Do

These templates generate candidate subject lines — they don't predict open rates. ChatGPT has no access to your list's historical behavior, your sender reputation, or the competitive inbox context your subscribers see. A subject line that tests well in one segment may underperform in another.

The model also can't account for deliverability signals. Certain words and patterns (heavy punctuation, all-caps, excessive symbols) can affect spam filter scoring depending on your ESP and domain history. Run outputs through your platform's spam check before sending.