← All articles

Cold Email Subject Lines: What Actually Gets Opened

October 12, 2026 · 9 min read · by the 0effort team

TL;DR: A good cold email subject line is a short, plain, specific preview of what the email says, written for one reader. Judge it by reply rate and positive-reply rate, not opens, because Apple Mail Privacy Protection and security scanners inflate opens. Skip fake "Re:" tricks and clever templates: they can win the open and lose the reply.

Open the average listicle on cold email subject lines and you'll find forty "proven" templates, each with a percentage lift attached. We don't trust the percentages, and we don't much trust the templates either. The question isn't "which line gets the most opens". It's "which line gets the right person to open an email they'll want to answer".

Why can't you trust open rates to judge a cold email subject line anymore?

Because an open is no longer a person. Apple's Mail Privacy Protection, launched with iOS 15 in 2021, stops senders from using invisible pixels to collect information about the reader (Apple, June 2021). In practice a mail client can fetch your tracking pixel whether or not anyone looked at the message. Corporate security scanners often fetch images and links before the recipient sees the email, with the same effect.

A test judged on opens therefore measures machines plus some humans.

The "this style lifts opens by X%" claims mostly come from vendor datasets with unstated samples, so we won't repeat any. Every number below names its source or prints its assumption.

What job does a subject line actually do in a cold email, and what happens after the open?

It gets the right person to open and keep reading, and it's honest about what's inside. That's all. A subject line can't sell anything. It can only promise, and the first line of the email either keeps the promise or breaks it.

Think of the open as a loan: a few seconds of attention borrowed against a claim ("this is about your hiring plans"). If the first sentence doesn't pay it back, you've spent trust and probably earned a deletion, or a spam report. A vague line that everyone opens and no one answers isn't a win. A line with fewer opens but more replies per delivered email is the better line.

What do cold email subject lines that get replies have in common?

Three things, we think: they're specific, plausible and colleague-sounding.

Specific means the line names what the email is about: a hire, a launch, a tool, a process. "Quick question" could be from anyone about anything, which is why it became noise. Plausible means a busy person would believe a one-to-one email might carry this subject. "Your Q3 SDR hiring" is plausible; "Revolutionary growth opportunity" isn't. Colleague-sounding means no title case, no exclamation marks, no brand voice.

Notice what's missing: tricks. And the line has to match the first sentence. If the subject says "Your SDR hiring", the email opens on SDR hiring, not on who you are.

How long should a cold email subject line be, and does lowercase or personalization really help?

Short. Our rule of thumb is two to six words, usually under about 50 characters, readable whole in a phone notification. That's a judgment, not a measured optimum. Clients truncate at different points, so send yourself a test.

Lowercase is the popular hack: write it like an internal message and it stops looking like marketing. The logic is plausible, but we know of no trustworthy data behind it, and a polished branded email with a lowercase subject can read as a costume. Test it; don't assume.

Personalization needs a distinction. A first name in the subject is the oldest mail-merge tell, so we'd skip it. A specific, true fact about the prospect's company or role (a hire, a launch, a tool they use) is what we'd call personalization, and it works because it's information, not decoration. If you can't back the fact up in the first line, keep it out of the subject.

Which subject line patterns hurt you, and why do "clever" templates backfire?

Most of the damage comes from lines that promise more than the email delivers. The ones we'd drop:

• Fake threading. "Re:" or "Fwd:" on a first email implies a conversation that never happened. Google's sender guidelines say not to start subject lines with Re: or Fwd: unless the message is an actual reply or forward (Google, accessed 2026-10-12), and US law prohibits subject headings likely to mislead about a material fact (15 U.S.C. § 7704(a)(2)).
• Invented history. "Following up on our call" when there was no call.
• Fake urgency. "Last chance" for an offer that doesn't expire.

Why do clever templates underperform at scale? Because they're shared. A template in a popular roundup sits in thousands of sequences, so your recipients have seen the pattern and know it. A cute surprise in one inbox is a recognizable tell in the same inbox a month later. Plain specificity doesn't wear out; it describes a fact about this recipient.

On spam: you'll read that certain words ("free", "guarantee") send mail to junk. We wouldn't build a strategy on that. Filters weigh sender reputation, authentication, engagement and volume far more than any single word. If your mail lands in spam, start with the deliverability checklist and SPF, DKIM and DMARC, not your adjectives. A misleading line does hurt in one measurable way: complaints. Google tells bulk senders to keep the spam rate under 0.10% in Postmaster Tools and never reach 0.30% (Google, accessed 2026-10-12).

Legal rules differ by country; for EU and UK recipients, start with our GDPR guide. None of this is legal advice.

How should the subject line, preview text and first line work together?

As one message in three parts. Most cold email is plain text, so there's usually no separate preheader: the inbox shows the subject, then the first line of the body as the preview. Whatever you write first is your preview text.

That's why "Hope you're doing well" is expensive. It spends the preview on nothing. Let the subject name the topic, the first line name your reason for writing to this person, and the rest make one ask. If the subject says "Series B and SDR hiring", the preview might read "Congrats on the round. I saw three SDR roles open up."

Read subject plus first line as a pair. If a stranger couldn't tell why they got it, rewrite the pair first. Full email structure lives in our cold email guide.

What are good cold email subject line examples by situation?

Patterns, not scripts. Swap in your own facts, and use a pattern only when the fact is true.

Our piece on B2B buying signals ranks which triggers deserve a line of their own.

How do you A/B test subject lines when opens are unreliable?

Test against replies and count honestly. The primary metric is reply rate per delivered email. The one that matters more is positive-reply rate: replies showing interest, a referral or a meeting, divided by delivered emails. Report negative replies and unsubscribes too, so a line can't "win" by annoying people.

Hold everything else equal. Same body, same offer, same sender pool, random assignment, same send days. Change only the subject.

Then do the math before launch, because reply events are rare. As an illustration, with assumptions printed (two-sided test, 5% significance, 80% power): detecting a positive-reply rate moving from 2% to 3% takes roughly 3,800 delivered emails per variant. From 5% to 7%, about 2,200. From 5% to 10%, about 430. Run your own baseline.

That sets the rules. Test two variants, not six; each extra one needs its own full sample. Don't crown a winner on 40 replies across the whole test. If your list is small, test bold differences (specific against generic, trigger against problem) and skip comma-level tweaks. Don't stop the moment one side leads; that manufactures false winners.

How does an AI SDR write and test subject lines across thousands of prospects without them all sounding the same?

Sameness is the failure mode. One prompt and a thousand prospects gives you a thousand variations of one line: a template with extra steps.

0effort is an autonomous AI SDR that sources buyers and writes and sends every email and LinkedIn touch. We'd hold any such system, ours included, to three rules:

1. The subject comes from a fact about that prospect, not from a style. With no usable fact, a plain generic line beats a forced one.
2. The subject is checked against the first line. If the email doesn't open on what the subject promises, it shouldn't send.
3. Evaluation uses replies and positive replies per delivered email, never opens, on enough sends to mean something.

Rule 1 matters most: variety from different real facts keeps a campaign from looking mass-produced; variety from synonyms doesn't.

FAQ

What is a good open rate for cold emails today?

We won't give a number, and we'd distrust anyone who does. Machine fetches inflate opens, so published benchmarks mix them with real reads, and most come from one vendor's customers. Track reply and positive-reply rates per delivered email and compare against your own history.

Should cold email subject lines include the prospect's name or company?

A first name, usually not: it reads as a mail merge. A company name, sometimes, when it carries information, as in "New SDR roles at Acme".

Do emojis, question marks or "Re:" in a subject line help or hurt cold email?

A fake "Re:" hurts: Google says not to use Re: or Fwd: unless it's an actual reply or forward, and misleading subject headings are prohibited under CAN-SPAM. We'd skip emojis in B2B. A genuine question mirroring the email's ask is fine; test it.

Do certain words in a subject line send cold email to spam?

Not on their own. Filters weigh sender reputation, authentication, engagement and volume much more than any single word. A pushy or deceptive line can raise complaints, though, and complaints damage reputation. Fix authentication and list quality first.

Should follow-up emails keep the same subject line or use a new one?

Keep it when the follow-up continues the same point; the thread supplies context in one click. Use a new subject when you change the angle or return after a long gap. Timing and touch count are in our follow-up cadence guide.

How many subject line variants should I test, and how many sends does each need?

Two, rarely three. Each needs its own full sample: with a 2% positive-reply baseline and a hoped-for 3%, roughly 3,800 delivered emails per variant at 5% significance and 80% power. That's a textbook calculation, not your data. With a small list, test only large differences.