Email Subject Lines: Curiosity vs Clarity (Prompt Picker)

A language-learning app ran a win-back campaign to 40,000 users who had gone quiet after breaking a daily streak. The team tested two subject lines on the same offer. The curious one, “Something changed while you were away,” beat the clear one on opens, but the clear one, “Your Spanish streak is gone. Restart in 5 minutes,” won on clicks and on the number of lessons started. Both were well written. They were solving different problems, and the team had never decided which problem the email was for. Choosing a subject angle is a prompt decision, and clarity about that choice is what makes the test readable.

This guide compares six subject-line angles you can hand to ChatGPT, with a decision table, the moment each fits, a copy-paste prompt for each, and the cost it carries. Every example stays inside one campaign: a win-back email to lapsed learners of a language-learning app. For cold outreach, where the reader has no prior relationship, the same angles behave differently, as covered in the cold email ChatGPT prompt. If you are weighing voice across a whole sequence, AI newsletter prompts covers the adjacent case.

Subject Line Trade-Off Matrix

Two questions pick the angle. First, does the reader already know they want the thing you are selling? A warm learner who just lapsed needs clarity, not intrigue; a dormant reader who has forgotten the app may need a hook to open at all. Second, what does success mean for this send: opens, clicks, or lessons started? Optimize the angle for the metric that matters, because curiosity and clarity rarely win the same one.

Choosing by Recipient State

Curiosity gap for cold attention. A subject such as “Something changed while you were away” works when the reader has no active goal and you must earn the first glance. It withholds the payoff on purpose. The risk is real: curiosity inflates opens while deflating clicks, so only use it when the email body delivers a payoff worth the tease. Never let the gap outlive the open; a bait-and-switch trains readers to distrust your name.

Direct benefit for a reader who is ready. “Restart your Spanish streak in 5 minutes” states the offer and the effort. A learner who already wants to resume does not need mystery, they need permission and a low-friction next step. The weakness is sameness; if every subject in your sequence is a benefit statement, they blur together, so vary the angle across a series rather than repeating one.

Specific number for habit and progress. “You were 3 lessons from a 30-day badge” is more persuasive than “come back today,” because it makes the loss concrete and the gain small. Numbers work best when they are true and tied to the reader’s own history. A hollow number, such as “join 10 million learners,” is wallpaper, and readers skip it.

Question for a genuine fork. “Still learning, or did life get in the way?” invites a self-assessment and can qualify a segment. It fails when the reader can answer with a single word, because a question they can dismiss in their head gives them no reason to open. Ask a question the reader cannot resolve without reading on.

Personalized token with clean data. “Ravi, your Spanish review queue has 42 cards” outperforms any generic line when the token is accurate. The entire edge depends on data quality: a wrong name or a stale number is worse than no personalization, because it proves the sender does not actually know the reader. Validate the merge fields before you send, every time.

Deadline for a real limit. “Your bonus streak freeze expires tonight” creates urgency that only works if it is true and if the thing at stake matters. Artificial scarcity, reset every week, teaches readers that your deadlines are noise. Use it sparingly, and only when the clock is genuine.

Prompt Templates by Angle

Curiosity gap

Write 8 subject lines for a win-back email to lapsed users of a language-learning app.
Angle: curiosity gap.
Rules:
- Withhold the payoff but hint that something meaningful changed.
- No clickbait that the email body cannot honor.
- Keep each line under 45 characters.
- Vary the wording so no two lines open the same way.
- After the list, note which line risks the weakest click-to-open follow-through and why.

Direct benefit

Write 8 subject lines for a win-back email to lapsed learners of a language-learning app.
Angle: direct benefit.
Rules:
- State the concrete outcome and the small effort required.
- Use plain verbs: restart, review, practice.
- Include the estimated time in at least three lines.
- Keep each line under 50 characters.
- Return a numbered list, no commentary.

Specific number

Write 8 subject lines that use one specific number from the reader's own history, for a win-back email from a language-learning app.
Placeholders available: {{first_name}}, {{days_since_last_lesson}}, {{cards_due}}, {{streak_length}}.
Rules:
- Each line uses exactly one placeholder.
- The number must create a small, achievable goal, not shame.
- Keep each line under 50 characters.
- Flag any line that would read as guilt-inducing if the number is large.

Question, personalization, and deadline

Write 6 subject lines for a win-back email from a language-learning app, two lines per angle.
Angles: question, personalized token, deadline.
Rules:
- Questions must not be answerable with yes or no alone.
- Personalization may use {{first_name}} only; do not invent facts.
- Deadlines must reference a real, near-term date field, not a fake reset.
- Keep every line under 50 characters.
- Label each line with its angle.

What You Give Up

Curiosity buys attention and charges you clicks. Clarity buys trust and charges you novelty. Numbers and personalization buy relevance and charge you data hygiene, because every merge field you add is a failure mode if it is empty or wrong. Questions buy engagement and charge you clarity when the question invites a one-word dismissal. Deadlines buy urgency and charge you credibility the moment the clock turns out to be decorative.

The practical method is to decide the metric before the words. If the goal is to wake a dormant list, lead the campaign with a curiosity or number angle and accept softer clicks on the first send. If the goal is conversions from a warm list, lead with benefit and clarity and expect a lower open rate with a higher action rate. Then test one variable at a time: same body, same offer, two angles, and read the result on the metric you chose, not on the one that happens to look better.

Whichever angle wins, keep a small library of the lines that worked so the next campaign starts from evidence instead of a blank box. Ask the model to group its output by angle and to flag the lines that risk overpromising, so the risky options are visible rather than shipped by default. A subject line is a promise the email must keep; pick the angle that lets you keep it, and the metric follows.

Sources & Notes

Prompts tested with ChatGPT on the GPT-4o and GPT-4.1 models in October 2026; the same angle instructions work in Claude and Gemini. The language-learning app, its win-back campaign, and the 40,000-user figure are an illustrative scenario, not a real product or send. Email marketing benchmarks vary widely by list and industry, so treat any open and click patterns here as directional, not normative. All prompts are copy-paste text; no external services were called. Category mapping follows the guunn.com AI Prompts plan (Marketing = 959).