Cold Email Subject Lines That Actually Get Opened

A cold email has one job before it has any other job: it has to get opened. Everything you put into the body — the research, the personalisation, the carefully chosen call to action — is worth precisely nothing if the subject line does not survive the two seconds your prospect spends scanning an inbox they did not ask you to be in.

And yet the subject line is usually the last thing anyone writes. It gets bolted on in a hurry, borrowed from a blog post listing "47 subject lines that convert", and fired at two thousand people at once. Then the open rate comes back at 14% and everyone blames the list.

This piece is about what actually determines whether a cold email gets opened. Not tricks — the tricks mostly stopped working years ago, and several of them now actively hurt you. What follows is the reasoning: what your reader sees, what the filter sees, and how to tell the difference between a subject line that works and one that merely sounds clever in a meeting.

The half-second decision your subject line is actually competing in

Picture the moment honestly. Your prospect is on a phone, between meetings, looking at a list. They are not reading. They are triaging. In that list, each email is roughly three pieces of information: who it is from, what the subject says, and — depending on the client — the first few words of the body.

They are not asking "is this interesting?" They are asking a much colder question: is this for me, and is it going to cost me anything to find out?

That is the entire decision. Not curiosity, not cleverness — relevance and cost. A subject line that signals "this is about your specific situation" wins. A subject line that signals "this is a mass email and I am about to be sold to" loses, regardless of how well written it is.

This is why the two most common failure modes are opposites of each other, and why fixing one often causes the other.

Failure one: the subject line that sounds like marketing

"Transform your pipeline with AI-powered lead generation." "Unlock 3x more meetings this quarter." "Quick question about your growth strategy."

These read as advertising, and people have been trained by twenty years of inbox exposure to delete advertising without thinking. The giveaway is usually one of three things: a benefit claim, a superlative, or a word that no human being uses in a one-to-one email. Nobody writing to a colleague says "unlock", "transform", "supercharge", or "revolutionise".

Worse, these are also the phrases that spam filters have the most training data on. You are simultaneously telling the human "I am selling" and telling the filter "I am bulk". You lose twice.

The fix is not to write a better benefit claim. The fix is to stop making a claim in the subject line at all. The subject line's job is to earn the open. The body's job is to make the case. Collapsing the two is what produces the marketing voice in the first place.

Failure two: the subject line that is clever for its own sake

The reaction against marketing language usually swings too far. You end up with "so this is awkward…" or "you're not going to like this" or a single lowercase word like "thoughts?".

These generate opens. They also generate resentment, because the open was obtained under false pretences. The reader clicks expecting one thing and finds a sales pitch. That gap — between what the subject promised and what the body delivered — is where reply rates go to die, and it is where spam complaints come from.

The metric to watch here is not open rate. It is reply rate divided by open rate. A curiosity-gap subject line will inflate the numerator of your open rate and quietly destroy the ratio that actually matters. If you are optimising for opens alone, you will optimise your way into a worse campaign while the dashboard tells you that things are improving.

What actually works: specificity that only applies to one person

The subject lines that consistently outperform have one property in common. They could not have been sent to anyone else.

Not "Quick question about your hiring" but "the three SDR roles on your careers page". Not "Improving your deliverability" but "your SPF record lists two providers". Not "Partnership opportunity" but "the integration you mentioned at SaaStock".

The mechanism is simple. Specificity is evidence of effort, and effort is the one thing a mass email cannot fake. When a reader sees a detail that could only have come from someone actually looking at their business, the "this is bulk" heuristic fails to fire, and the email gets read on its merits.

This is also, incidentally, why the personalisation token approach — "Hi {{first_name}}, I saw {{company}} is hiring" — stopped working. A merge field is not specificity. Everyone knows what a merge field looks like. Real specificity requires you to have actually looked, which is why it scales badly, which is exactly why it works.

The length question, and why the usual advice is wrong

You will read everywhere that subject lines should be short. Six words. Four words. Three.

The advice is right for the wrong reason, and the wrong reason leads people astray. Short subject lines do not win because brevity is inherently persuasive. They win because mobile clients truncate, and because a short line is harder to fill with marketing language.

What matters is not the word count but where the truncation falls. Most mobile clients show somewhere between 35 and 45 characters — Litmus keeps a running chart of what each client actually displays. Anything after that is invisible on the device where most of your email is triaged. So the real rule is not "be short" — it is put the load-bearing word in the first 35 characters.

"Question about the three SDR roles you're hiring for" is 51 characters, and it works fine, because the specific detail lands before the cut. "Following up on my previous email about your SDR hiring" is shorter in spirit and completely useless, because the first 35 characters contain nothing at all.

Write the line, then look at the first 35 characters in isolation. If those alone would not earn the open, the rest of the line is decoration.

What the spam filter sees, and why it is not what you think

There is a persistent myth that spam filters work from a list of banned words — that writing "free" or "guarantee" will send you to the junk folder. This has not been meaningfully true for a long time. Modern filtering is overwhelmingly reputation-based, and content is a small and shrinking input. Microsoft says as much in its anti-spam documentation.

What actually determines placement, roughly in order of weight:

• Your sending domain's history. Have people opened, replied, and not complained about mail from this domain before?
• Authentication. Does SPF, DKIM and DMARC line up? If not, you are starting from a deficit that no subject line can recover. You can check yours with our free SPF, DKIM and DMARC validator.
• Recipient engagement. Bounces, deletions-without-opening, and "mark as spam" all feed back into the reputation of the domain that sent them.
• List quality. Sending to dead addresses is one of the strongest negative signals there is, which is the subject of the next section.
• Content. A long way down. It matters, but far less than the four things above.

Google is unusually direct about this in its sender guidelines, which are worth reading in full if you send any volume of cold email at all. The short version: they care about whether recipients want your mail, and they measure that behaviourally.

The practical implication for subject lines is liberating. You do not need to contort your language to dodge a filter. You need a clean list, correct authentication, and a subject line that makes real humans want to open — because it is the humans' behaviour that the filter is ultimately reading.

The bounce rate problem that no subject line can fix

Here is the failure that masquerades as a subject line problem. Open rates collapse. Somebody rewrites the subject lines. Open rates do not recover. Somebody rewrites them again.

The actual cause is often that a meaningful share of the list does not exist. Every send to a dead address is a bounce, and a high bounce rate does two things at once: it removes people from the denominator who were never going to open anything, and it damages the sending reputation that determines whether the valid addresses see your mail at all.

A list with 20% dead addresses does not cost you 20% of your results. It costs you considerably more, because the reputation damage suppresses delivery to the 80% who are real. This is the single most common reason a campaign underperforms for reasons that look like copy and are not.

Verify before you send. Our free email verifier will check syntax, domain, and mail-server records for a single address, and Govarova's verification API does it at volume, including catching the disposable addresses and spam traps that do the most reputational damage. If you have never cleaned a list, the first clean is usually the largest single improvement available to you — larger than any rewrite.

Six subject-line patterns that hold up

These are not templates to copy. They are shapes, and the specificity has to be yours.

1. The observed detail. "the two SDR roles on your careers page". You looked. That is the whole message.
2. The mutual reference. "Sarah suggested I reach out". Only if true. This one is powerful precisely because it is verifiable, which means abusing it is unusually costly.
3. The specific problem. "your signup form accepts disposable emails". A statement of fact about their business, which they can check in ten seconds.
4. The plain question. "who owns outbound at [company]?" Low effort to read, low effort to answer, and it does not pretend to be anything other than what it is.
5. The relevant event. "your Series A" or "the London office". Something that happened, that you know about, that changes their situation.
6. The direct ask. "15 minutes on [specific topic]?" Honest about the intent. Underrated, because it costs the reader nothing to evaluate.

Notice what none of these do: none make a claim, none use a superlative, and none manufacture curiosity. Each one either demonstrates that you looked, or asks something plainly. That is the entire repertoire, and it is enough.

If you want a starting bank of lines to react against, our free cold email subject line generator will produce and score a set for a given topic — useful less as a source of finished copy than as a way of seeing several angles quickly and noticing which ones you instinctively reject.

How to test a subject line without burning your list

The instinct is to A/B test. The problem is that most cold email lists are far too small for an A/B test to tell you anything. To detect a five-point difference in open rate with any confidence, you need hundreds of sends per variant. Split a list of 400 in half and the result you get is noise wearing the costume of data.

What works better at small volumes:

• Send in waves, not batches. Take the first 50, send, read the results, adjust, send the next 50. You are not running an experiment; you are iterating. That is the correct method when your sample is small.
• Read the replies, not just the rates. Fifty sends will produce a handful of replies, and those replies contain more information than any open-rate delta. "Not the right person" is a targeting problem. "We already have one" is a positioning problem. Neither is a subject line problem.
• Watch the ratio, not the rate. As above: replies per open. A subject line that raises opens and lowers that ratio has made your campaign worse.

If you do have the volume for a real test, model it first. Our free open rate calculator will show you what a given lift is actually worth against your list size, which usually reveals that the test you were about to run could not have detected the difference you were hoping for.

Why "Re:" and "Fwd:" are a bad idea, and what they cost you

Someone will eventually suggest prefixing your subject line with "Re:" to make it look like a reply to a conversation that never happened. It works, in the narrow sense that it raises open rates.

It is also a lie, and it is a lie that the reader detects within one second of opening — the moment they scan the thread and realise there is no thread. Everything that follows is read by someone who now knows you are willing to deceive them to get attention. You have spent your credibility before you have made your argument.

There is a commercial argument against it too, quite apart from the ethical one. Mailbox providers are alert to fake threading, and a spike in opens followed by immediate deletions is a signature they read as manipulation. The behaviour you are producing — opened, then deleted in under three seconds, then eventually reported — is close to the definitional pattern of unwanted mail.

The same goes for fake "Fwd:", for manufactured urgency ("URGENT"), and for the "as discussed" opener when nothing was discussed. Each one buys a short-term open at the cost of the relationship you were ostensibly trying to start. If your subject line only works because the reader has been misled about what it is, you do not have a subject line. You have a trap, and people remember traps.

The sender name is doing more work than the subject line

Here is something that consistently surprises people who spend hours on subject lines: the "from" field is often the stronger signal.

The reader's first question — is this for me? — is frequently answered before they read a single word of the subject. A message from "Sarah Chen" is evaluated differently from one from "Govarova Marketing Team", which is evaluated differently again from "noreply@". The first is a person writing to a person. The last is a machine, and it will be treated as one.

Three practical consequences follow:

• Send from a human, with a real name. A named sender with a real, monitored mailbox will outperform a brand alias almost every time in a cold context. Brand names work for newsletters people opted into; they work against you in a first contact.
• Make the reply address real. If someone replies and it bounces, you have converted your best possible outcome into a negative signal. It happens more often than anyone would like to admit.
• Keep the sending identity stable. Reputation accrues to a domain and an address. Rotating sender names to dodge filters resets that accrual and looks, to the receiving server, exactly like evasion — because it is.

You cannot fix a bad sender identity with a good subject line, and a strong sender identity forgives a fairly ordinary one.

Preview text: the half of the subject line nobody writes

Most clients show a snippet of the body next to the subject — 40 to 90 characters of it. That snippet is part of the subject line as far as the reader is concerned, and almost nobody writes it deliberately.

Which means it usually contains whatever happened to be at the top of your email: "Hi Sarah," or, worse, "View this email in your browser" or an image alt tag. You have been given a second line of free real estate directly beside the first, and you are spending it on the word "Hi".

Treat the first sentence of the body as a continuation of the subject, not as an opening pleasantry. If the subject is "the three SDR roles on your careers page", the preview should carry the thought forward — "you're hiring three at once, which usually means the ramp is the problem" — not reset it with a greeting.

Read the two together, as the reader will, in a list, on a phone. If they do not make sense as one thought, you have wasted half of your available space.

What to do when the open rate is fine and nothing else is

A campaign can have a healthy open rate and be completely broken, and this is the case most likely to waste a quarter.

Opens at 45%, replies at 0.3%. The instinct is to keep tuning the subject line, because that is the number that is visibly working. But the subject line has already done its job — people opened. Something after the open is failing, and no amount of further subject-line work can reach it.

Work backwards through the funnel instead:

• Opens high, replies near zero. The body is not landing, or the ask is too large, or the person is not the buyer. Look at who replies at all, and what they say.
• Opens high, unsubscribes or complaints high. The subject line is over-promising. You are winning the open dishonestly and paying for it immediately.
• Opens low, but replies-per-open strong. This is the good problem. Your message resonates with the people who read it; you simply need more of them to read it. This is the only situation in which subject-line work is genuinely the highest-value thing you can do.
• Everything low, and bounces above 3%. Stop. This is not a copy problem. Your reputation is being damaged with every send, and the fix is in the list — see our piece on email verification and sender reputation.

Diagnosing which of these four you are in takes ten minutes and saves weeks of optimising the wrong thing. Most teams never do it, which is why most teams rewrite subject lines that were never the problem.

A worked example, start to finish

Take a real situation. You sell an email verification API. Your prospect is a fintech with an open signup form.

The lazy version: "Improve your signup quality with AI-powered email verification". It makes a claim, it uses "AI-powered", and it could have been sent to ten thousand companies. It will be deleted unread, and it deserves to be.

The curiosity version: "quick question". Higher open rate. The reader opens, finds a pitch, and feels mildly conned. Reply rate collapses, and the ratio that matters gets worse while the metric on the dashboard gets better.

The specific version: "your signup form accepted mailinator.com". You went to their form. You typed in a disposable address. It went through. That is a fact about their business, it took you ninety seconds to establish, and no competitor sending mass email has done it.

The preview text carries it forward: "took about ninety seconds to test — happy to send you what came back." The body is three sentences and one ask.

The third version cannot be sent to ten thousand people, and that is not a weakness of the approach. It is the entire point. Fifty of those beat ten thousand of the first, and they beat them on reply rate, on reputation, and on the quality of the conversations that result.

If you want to see what such a check actually returns, run a disposable address through our disposable email checker — that output, pasted into an email, is a more persuasive subject line than anything you could have written.

The subject line is downstream of the list

Here is the uncomfortable conclusion, and it is the one worth taking away.

Most subject-line problems are targeting problems. If you are writing to people who have no plausible need for what you sell, there is no sequence of words that fixes that. You can raise the open rate — curiosity gaps will do it — and you will simply have more people opening an email that was never relevant to them, which produces deletions, complaints, and a slowly poisoned sending domain.

The highest-leverage improvement to your subject lines is almost always made before you write them: send to fewer, better-chosen people, about something you can demonstrate you actually looked into. A specific subject line is only possible when you have something specific to say, and you only have something specific to say when the list is right.

That is why the sequence matters. Build the list properly — enrich it so you know who these people are, verify it so you are not burning your reputation on dead addresses — and the subject lines will start writing themselves, because there will finally be something true and particular to put in them.

If you want the whole thing in one place, our guide to AI-powered lead generation covers the pipeline from finding the right people through to the sequence that reaches them. The subject line is the last five per cent of that work, and it only pays off when the other ninety-five is done.

Subject lines only earn the open; the rest of the sequence earns the reply. Govarova Sequences runs the whole cadence on verified data, and you can start free.