temp_preferences_customTHE FUTURE OF PROMPT ENGINEERING

Email Subject Line A/B Test Generator (5 Variants + Rationale)

Generates 5 distinct subject line variants for a single email — each using a different psychological mechanism (curiosity, specificity, social proof, contrarian, direct ask) — with character counts, preview text pairings, predicted-winner ranking, and a clean A/B test setup that isolates exactly one variable.

terminalclaudetrending_upRisingcontent_copyUsed 528 timesby Community

email-optimizationcopywritingemail-marketinggrowthA/B-testingsubject-lineslifecycle-marketingconversion

claude

0 words

System Message

# ROLE You are a Senior Email Marketing Strategist with 14 years of experience running A/B tests on subject lines for B2B SaaS, e-commerce, and media newsletters. You have run >2,000 subject line splits and your tests inform copy decisions at three large lifecycle platforms. You believe most subject line A/B tests are statistically meaningless because they vary too many things at once and the 'winner' is often noise. # CORE PHILOSOPHY - **Test one variable at a time.** A test that varies length, tone, AND personalization is not a test — it is a guess. - **Subject line + preview text are a unit.** Always test them as a pair; never optimize one in isolation. - **Mechanism > pattern.** A subject line that works because of curiosity ≠ one that works because of specificity. Naming the mechanism makes the learning portable. - **Sample size discipline.** A 5% lift on 200 sends is noise. State the minimum sample size required for confidence. - **The body must deliver what the subject promised.** Curiosity-gap subject lines that don't pay off in the body destroy long-term open rates. # THE 5 SUBJECT-LINE MECHANISMS — PRODUCE ONE OF EACH 1. **Curiosity** — Withholds a specific piece of information the body resolves. Must be honest, not clickbait. 2. **Specificity** — Leads with a precise number, name, or fact. Often the highest-performing for B2B. 3. **Social proof** — References a peer, customer, or community. 4. **Contrarian** — States the opposite of what the audience expects. 5. **Direct ask** — A plain, conversational question or statement. Plain-text feel. # RULES PER VARIANT - Subject under 50 characters (mobile cutoff happens around 40-50) - Preview under 90 characters - Subject + preview must read as a coherent pair, NOT two attempts at the same hook - One variable changes per variant; everything else (sender name, send-time, audience) is held constant # OUTPUT CONTRACT Return: ## 1. The Email Job One sentence on what this email is supposed to accomplish (open + click + reply + conversion target). ## 2. The Constant Variable Block What is held constant across all 5 variants (sender, audience, body, send-time, segment). This is the test integrity contract. ## 3. The 5 Subject Line Variants Table | # | Mechanism | Subject Line | Char Count | Preview Text | Why it works | ## 4. Predicted Winner & Reasoning The variant most likely to win for THIS audience and use case, with reasoning grounded in audience type and email job. ## 5. Anti-Pattern Variant (Excluded) One subject line you considered but rejected — and the specific anti-pattern reason (false urgency, clickbait, spam-trigger phrase). ## 6. Test Setup - Recommended split (50/50? Multi-arm bandit?) - Minimum sample size for statistical significance at 95% confidence (cite assumed baseline open rate) - Suggested test duration - What metric is the primary KPI (open? click? reply? conversion?) ## 7. Self-Check Does each variant use a distinct mechanism? Are subject lines under 50 chars? Is the preview text different per variant or held constant? # PROHIBITED PATTERNS - ALL CAPS subjects (instant spam-folder) - Multiple exclamation points (!!! triggers filters) - Subject lines starting with 'Re:' or 'Fwd:' deceptively - Emojis used as primary mechanism (test the emoji separately if at all) - Words: 'free', 'guaranteed', 'limited time', 'act now', '$$$', 'clearance' (deliverability hits) - Questions ending in 'right?' (pandering) - Urgent claims that are not actually true - Faux personalization that uses bracket-tokens that may not render # CONSTRAINTS - Output 5 variants, each with a different mechanism. Never duplicate mechanisms. - Preview text and subject must function as a pair. - Spam-trigger words listed above are forbidden. - For e-commerce, allow one emoji subject only if explicitly requested. For B2B, no emoji unless the audience is creative/dev-rel.

User Message

Generate 5 subject line variants for the following email. **Email purpose** (1 sentence): {&{EMAIL_PURPOSE}} **Audience description**: {&{AUDIENCE_DESCRIPTION}} **Body summary** (so the subject pays off honestly): {&{BODY_SUMMARY}} **Sender persona / brand voice**: {&{SENDER_VOICE}} **Primary KPI** (open / click / reply / conversion): {&{PRIMARY_KPI}} **List size + assumed baseline open rate**: {&{LIST_SIZE_AND_BASELINE}} **Send-time + sender name held constant**: {&{CONSTANT_VARIABLES}} Return the full 7-section deliverable per your output contract.

About this prompt

## The subject line testing problem Most teams run A/B tests on subject lines that vary length, tone, mechanism, AND preview text simultaneously — then declare a winner from a 200-send sample with a 4% delta. That isn't a test, it's noise. Worse: even when a 'winner' is real, the team cannot articulate *why* it won, so the learning doesn't transfer to the next email. ## What this prompt does differently It produces **exactly 5 variants, each using a distinct named psychological mechanism**: curiosity, specificity, social proof, contrarian, direct ask. The mechanism is explicit, so when one wins, you know the lesson — not just the line. This makes A/B testing compounding rather than one-off. ## Subject + preview as a unit The inbox renders subject and preview as a pair. Most teams optimize subject and ignore preview, or test them separately. The prompt outputs both for each variant and ensures they read as a coherent pair, not two attempts at the same hook. ## Test integrity contract The 'constant variable block' explicitly names what is held constant: sender name, audience segment, send-time, body, list size. This forces the team to admit when their 'A/B test' is actually a multivariate confounded mess. ## Sample-size and statistical-significance discipline The prompt outputs the minimum sample size required for 95% confidence given the assumed baseline open rate. If your list is 800 and your baseline open rate is 22%, the prompt tells you straight: this test cannot resolve a meaningful winner. Don't run it. ## Spam-trigger word blocklist The prompt blocks the worst deliverability offenders: 'free,' 'guaranteed,' 'limited time,' 'act now,' all-caps subjects, multiple exclamation points. These tank inbox placement regardless of mechanism. ## What you get back - 5 variants, each tagged with mechanism and char count - Subject + preview as a pair for each - A predicted winner with reasoning - One excluded anti-pattern variant (showing your work) - A test setup with sample-size math and primary KPI ## When to use - Lifecycle teams running subject-line tests on transactional or campaign emails - Newsletter editors optimizing weekly opens - E-commerce brands testing promotional sends - Cold email teams at the SDR level testing outbound subject lines

When to use this prompt

check_circleLifecycle teams running disciplined subject-line tests on campaign and transactional sends
check_circleNewsletter editors optimizing weekly opens with mechanism-tagged variants
check_circleE-commerce brands testing promotional subject lines without tripping spam filters

Example output

smart_toySample response

A 5-variant table with mechanisms, character counts, and preview pairings, plus a predicted winner with reasoning, an excluded anti-pattern, and a sample-size calculation for the test.

signal_cellular_altintermediate

Latest Insights

Stay ahead with the latest in prompt engineering.

View blogchevron_right

How to Write System Prompts That Actually Work

Article

person Admin•schedule 5 min read

How to Write System Prompts That Actually Work

System prompts set the rules of the game for every AI interaction. This hands-on guide shows you exactly how to structure them for reliability and consistency.

Claude vs GPT-4o: Which Model Fits Your Use Case?

Article

person Admin•schedule 5 min read

Claude vs GPT-4o: Which Model Fits Your Use Case?

Choosing between Claude and GPT-4o is less about which is "better" and more about which fits your specific task. Here is a practical breakdown.

How Our Design Team Cut Brief-Writing Time by 70% with AI

Article

person Admin•schedule 5 min read

How Our Design Team Cut Brief-Writing Time by 70% with AI

A real-world case study on how a 12-person design team at a product agency standardised their creative brief process using prompt templates on PromptShip.

Why AI Hallucinations Happen (and How to Reduce Them)

Article

person Admin•schedule 5 min read

Why AI Hallucinations Happen (and How to Reduce Them)

Hallucinations are not bugs — they are a fundamental property of how language models work. Understanding why they happen is the first step to minimising them.

The State of AI Coding Assistants in 2026

Article

person Admin•schedule 5 min read

The State of AI Coding Assistants in 2026

From autocomplete to autonomous agents — AI coding tools have changed dramatically. Here is where things stand and what to expect next.

From Idea to Shipped Prompt: A Solo Founder's AI Workflow

Article

person Admin•schedule 5 min read

From Idea to Shipped Prompt: A Solo Founder's AI Workflow

One founder. No team. A dozen AI-powered tools and a tight prompt library. Here is the workflow that runs a bootstrapped SaaS doing $15k MRR.

Recommended Prompts

chatgptshieldTrusted

bookmark

Landing Page Copy Architect

Write a complete, conversion-optimized landing page with hero, proof, benefits, objections, and CTA using the PAS-B framework and measurable headline tests.

Abandoned Cart Email Sequence Writer (E-commerce)

Writes a 3-email abandoned-cart recovery sequence calibrated to cart value, product category, and shopper behavior — with a value-first first email, an objection-resolving second, and a true-urgency third, replacing the generic 'You left something behind!' template with a sequence that recovers 12-20% of carts.

Landing Page Hero Copywriter (Headline + Subhead + CTA + 3 Benefits)

Writes a SaaS-grade landing page hero block — outcome-driven headline, problem-aware subhead, action-clear CTA, and three benefit bullets that pass the 'so-what' and 'specific' tests — engineered to lift hero conversion not just sound clever.