How to write email subject lines that get opened

Good email subject lines earn a click from someone skimming an inbox in under a second. The formulas that win are short (under 50 characters, with the hook inside the first 41 for mobile), specific, and focused on a single promise: curiosity, urgency, personalization, or a concrete number. Change one variable at a time and test the variants.

Subject lines still decide whether the rest of your email matters. Average open rates vary enormously by sector, so start from the industry benchmarks rather than a single headline figure; what stays constant is that on one list, the spread between a strong subject line and a weak one can reach 2–3x. That's a lot of leverage for one sentence.

What makes a subject line good

A great subject line promises something specific, sounds like a person wrote it, and gets to the point before the inbox truncates it. The pattern across the best-performing ones I've written: one clear benefit, a concrete noun or number, and a tone that matches the sender the reader already trusts.

The mechanism underneath all of that is relevance. A reader glancing at a crowded inbox is asking one question, which is whether this concerns them; a subject line that answers it earns the half-second needed to decide.

There's a short checklist I run before any campaign goes out. Does the line still work if the preheader gets cut? (Mobile clients sometimes drop preview text entirely.) Would I open it myself at 7am on a phone? Does it tell the truth about what's inside? If any answer is no, the line goes back.

How long a subject line should be

Aim for 30 to 50 characters, roughly six to ten words. Mobile inboxes cut off around 41 characters on iPhone in portrait and closer to 30 on some Android clients, so anything past that either wraps or gets an ellipsis. Short lines also read as more human; nobody texts a friend in 80-character sentences.

Here's the nuance that rarely gets mentioned. Word count matters less than character count on mobile, and character count matters less than where the hook sits. "Your March invoice is ready (and you overpaid)" is 46 characters, but the important word, "overpaid," lands at position 39, just inside the mobile window. Rewrite that as "Your March invoice is ready and we owe you money because you overpaid last cycle" and the hook is buried past truncation. Same information, very different open rate.

To see exactly how a line will land before you send it, run it through the subject line tester; it renders your subject and preheader the way Gmail, Apple Mail, and Outlook will display them.

The patterns that consistently outperform

Five patterns beat the rest reliably: curiosity, specific numbers, benefit promises, personalization, and urgency when the urgency is real. Matching the pattern to the audience is the whole trick. Curiosity lines fatigue B2B lists quickly; urgency burns out consumer lists if you lean on it weekly.

Curiosity works when the gap between what the reader knows and what they want to know feels small enough to close with one click. "What 312 SaaS founders told us about churn" promises a quick payoff. "You won't believe what happened next" promises nothing and smells like bait. The curiosity subject line collection is a useful calibration exercise for where that line sits.

Numbers do most of their work through specificity. "5 ways to cut onboarding churn" outperforms "ways to cut onboarding churn" almost every time, because the reader can price the commitment before opening.

Benefit promises have to survive the so-what test, the reader's instinctive "why should I care?" "Save 3 hours on your next campaign" beats "New Mailneo automation features" because the first is about the reader. Name the thing they save, earn, or avoid.

Personalization goes well beyond {{first_name}}, which is the low end of the range. The high end inserts behavior: the plan they're on, the last template they used, the city they're in, the feature they clicked last week. The email personalization guide covers the mechanics, and the personalization examples show what behavioral tokens look like in a live subject line.

Urgency has a short shelf life. "Ends tonight at midnight" works once. "Last chance!" three times a week teaches your list to ignore you. A closing window, a real deadline, or genuinely limited inventory pulls clicks; fake urgency trains people to unsubscribe, and you can watch it happen in your unsubscribe rate. If you're writing to a deadline, the urgency examples are worth a scan for phrasing that doesn't shout.

What personalization actually changes

Name-only personalization gives a modest bump. Behavioral personalization gives a much bigger one, because it proves you know the reader rather than their CRM record. "Sarah, we noticed you left something in your cart" works; "Sarah, it's been 43 days since you last logged in" reads like a surveillance report.

A few rules I've landed on after too many reply-all disasters:

• Default every token. If {{first_name}} fires on an empty field, the reader sees "Hi , here's your…" and your open rate tanks. Every template needs a fallback, even if it's just "there".
• Don't personalize on stale data. "Loved your last order of [item from 2022]?" in 2026 reads as negligence.
• Try the personal token in the preheader instead. On two Mailneo accounts I've run, opens went up when the token moved to preview text and the subject stayed a clean benefit promise.

Testing subject lines without fooling yourself

A good subject-line test changes one thing, sends each variant to a meaningful slice of the list (10–20% for large lists), and picks a winner on opens or clicks before rolling out to the rest. Anything looser is pattern-matching dressed as testing.

If A says "5 ways to save on shipping" and B says "Your order just got faster", you're testing curiosity against urgency, plus length, plus tone, all at once. You'll get a winner; you won't get a reason. One variable per test, then change the next variable next time.

On sample size: a 10% split of a 2,000-person list gives you 200 recipients per variant, which is too small to trust unless the winning lift is enormous. I use 1,000 recipients per variant as a floor, or I pool results across several sends before I let a result become a rule. A false positive in subject-line testing is expensive in a quiet way; you keep writing bad lines while believing they're good.

Send both variants at the same time of day, ideally within the same hour. Testing A on Monday at 8am against B on Tuesday at 2pm gives you a day-of-week test with a subject-line change bolted onto it, and the two effects are then impossible to separate. (If day of week is what you actually want to know, start from the send time data and design that test on purpose.) The full statistical walkthrough lives in the A/B testing guide.

Then measure both halves. Open rate tells you the subject line worked; click rate tells you the email kept the promise. A line that lifts opens 15% and drops clicks 20% is a loss. Optimize for the downstream metric that pays the bills.

AI-generated variants feed A/B tests well, because a model reliably produces different angles on request: a curiosity version, a number-led version, a benefit version. See the AI email writing guide for the prompting approach. Don't ship the output raw; every high-performing line I've tested had at least one human edit, usually a cut word or a swapped verb.

The mistakes that tank open rates

Spam trigger words still hurt, though less through filters than through people. "FREE!!!", "Act now", "100% guaranteed", and "Make $$$" get flagged by readers who then mark you as spam, and complaints damage sender reputation far more directly than any keyword list ever did. Skip the shouting, the stacked exclamation points, and the money promises; the spam folder avoidance guide has the longer version.

Clickbait costs you twice. "You won't believe this" earns one click and a permanent trust hit, and when the body doesn't pay off the promise, unsubscribes spike and every future campaign inherits the damage.

Vagueness is the quiet killer. "An important update" tells the reader nothing. Neither does "Quick question" in a newsletter, although in B2B cold outreach it works surprisingly well because the context is different; the cold outreach subject lines show why. Every subject line should answer "about what?" in seven words or fewer.

All-caps reads as shouting, and inboxes have been training people to that reaction since the 1990s. Emoji are more nuanced: one, at the start or end, can help in consumer categories and consistently underperforms in B2B and financial services. The emoji subject line examples are worth checking against your own vertical before you commit.

Then there's the preheader, which does half the work and gets none of the attention. Subject and preview text operate as a pair. If the subject is short, the preheader extends it; if the subject asks a question, the preheader can start answering. Leave it blank and Gmail fills it with the first line of your email, which is almost always "View this email in your browser." Check it with the email preheader previewer.

Finally, don't lean on one pattern forever. If every subject line is a question, your list tunes out. If every line uses urgency, urgency stops meaning anything. Rotate across a campaign series and the inbox stays unpredictable.

Templates worth starting from

These are the patterns I return to when writing from scratch. They work across verticals when the content genuinely delivers on them, and the tone still has to match your brand.

For fully written lines rather than skeletons, the category collections are quicker to raid: ecommerce, SaaS, newsletter, welcome, abandoned cart, and re-engagement.

One honest limitation: templates get you to a baseline and no further. The lines that outperform everything else sound like a specific person writing to a specific reader. Use the table as training wheels and drop it once your list starts recognizing you.