Articles / Cold Email Subjects Are a Filter, Not a Hook

cold email

Cold Email Subjects Are a Filter, Not a Hook

Finn ·

A cold email subject is the shortest true description of the one thing you are asking for. Not a hook, not a curiosity gap: a label. Its job is to filter, so the people who open are the people who could plausibly say yes, and everyone else deletes it from the list view without ever costing you anything.

The definition that circulates

Open any guide on cold email subjects and you find the same working definition: the subject line is a hook, and its job is to maximize the open rate. Everything else follows from there. The curiosity gap. The lowercase three-word line engineered to look like it came from a colleague. "Quick question." The first name dropped in to trigger recognition. The lists of twenty or fifty tested lines you are invited to swipe.

The reasoning holds as far as it goes: nobody replies to an email they never opened. What breaks is the step after, where the open stops being the gate and quietly becomes the goal.

What the hook definition misses

Opens are not scarce. You can buy them. A vague, intriguing line lifts open rate on almost any list, because curiosity is cheap to trigger and expensive to satisfy. The reader opens, finds a pitch with no relation to the promise, and closes it. You converted an open into a delete, plus a little resentment, and on a bad day into a spam complaint that taxes the next few hundred sends.

The second thing the hook framing misses is that you are not writing one field. You are writing three. An inbox row shows the sender name, the subject, and a snippet of the body, and the reader takes them in as a single line. Only the middle one feels like the subject, yet the snippet is simply your opening sentence. Which means "Hi Sarah, hope this email finds you well" spends a third of your inbox real estate saying nothing at all. Rewrite that first sentence and you have extended your subject line without touching it.

The third miss is legal, not stylistic. "Re:" on an email that answers nothing and "Fwd:" on a thread that never existed are deceptive, and under US CAN-SPAM rules a deceptive subject line is a violation rather than a tactic. A merge tag firing on an empty field fails the same way: it does not read as personal, it reads as a mail merge caught mid-failure.

The subject line is a test on the body, and no swipe file fixes an email that fails it.

A cold email subject, defined properly

A cold email subject is the shortest true description of the single thing you are asking for, written after the body, and understandable by someone who has never heard your name.

Three properties carry the definition. True, meaning the body delivers exactly what the subject announced, with no gap left to close. Single, meaning one ask, therefore one subject. Short, because a phone shows roughly the first 35 characters and everything past that is only visible on a desktop client.

Then comes the part that changes how you write. Try to compress your email into six words and one of two things happens. Either the ask lands cleanly, or it refuses to fit, and the reason it refuses is almost never the wording. It is that the email is asking for two things, or asking for nothing, or dressing up a demo request as a favor. The subject line is a test on the body, and no swipe file fixes an email that fails it.

That is why the subject gets written last. You cannot summarize an ask you have not settled yet, and the whole point of a template like the Tim Ferriss cold email template is to force the ask into a shape small enough to summarize.

The subjects this definition produces

Applied to real asks, the rule produces lines like these:

  • Asking for 20 minutes about a specific problem: 20 minutes on your onboarding emails
  • Asking for an introduction: intro to your ops lead?
  • Asking someone to look at a thing you built: a scheduling tool for photographers, 2 minutes
  • Sending something useful with no ask attached: teardown of your pricing page, no strings
  • Following up on a change you noticed: your new checkout flow

They read plainly, and that is the point. Each one is a label a stranger decodes in the list view, each one is paid back by the first line of the body, and each one names something too specific to have been sent to a thousand people. Specificity is the personalization. The first name never was.

The lines you should be suspicious of are the ones that would work equally well on any recipient in your list. "Quick question" is the clearest case. It is true of every email ever sent, so it filters nobody, and after a decade of overuse it reads as the signature of a sequence rather than of a person.

Measuring subjects when opens lie

Since Apple introduced Mail Privacy Protection with iOS 15 in September 2021, Apple Mail pre-loads tracked images whether or not the message is read. A large slice of your open rate now measures a proxy server, not a human. Optimizing subject lines against that number is measuring the instrument.

For a solo sender the practical answer is to stop A/B testing subjects altogether. At 40 or 60 sends a week you will never reach a result worth trusting, and the variance between two subject lines is smaller than the variance between two lists. Group your subjects into families instead, three at most: descriptive ask, direct question, referral or context line. Send one family for a few weeks, count replies per hundred sends, then switch. You are looking for a difference big enough to see without statistics, and if there is none, the subject was never your constraint.

Usually it is not. The ceiling is set upstream, by whether the domain is trusted enough to land at all, which is the entire business of a service like Mailivery, and downstream by what you do with silence. An opened email that goes unanswered is not a subject line problem; it is the moment the follow up rule decides the outcome, one new fact per message rather than a polite reminder that time has passed. And if you are tempted to hand the whole sequence to an AI agent, the same logic that applies to picking an OpenClaw alternative applies here: a narrow agent doing one job you can inspect, never a general one improvising your reputation.

So write the email, settle the ask, then compress it into six honest words. If the words will not come, do not reach for a better hook. Go back and fix the email.

FAQ

How long should a cold email subject line be? Aim for three to six words, or about 35 characters, because that is roughly what a phone shows before truncating. Nothing penalizes a longer subject, but anything past the cut reaches desktop readers only, so never put the decisive word at the end.

Should I put the recipient's first name in the subject? Rarely. A first name signals a mail merge as often as it signals attention, and it costs characters you need for the ask. Naming something specific to the recipient, a product, a page, a recent change, does the job a first name is supposed to do.

Do lowercase subject lines get more replies? Lowercase reads as informal and can help when it matches how you actually write. It is a register choice, not a lever. A lowercase line on a generic email is still a generic email, and inconsistency between subject and body is what readers notice.

Is "quick question" still a good cold email subject? No. It is true of every cold email, so it filters nobody, and it is common enough that experienced recipients read it as sequence boilerplate. Replace it with the actual question, shortened.

Should the subject line be a question? A question works when it is the real ask and the reader can answer it in one word. It fails when it is rhetorical, or when the body then asks for something else, because that gap is what burns the reply you were trying to earn.

Save months.

Products, marketing, tools: I test a lot, sum up what works, you avoid my mistakes.

Read next

LinkedIn's Post Inspector, at linkedin.com/post-inspector, reads a URL the way LinkedIn's crawler does and clears the preview LinkedIn had stored for it, so your next share picks up the tags currently on the page. That is the whole job: it refreshes, it does not validate. When the preview is still wrong after you run it, the cache was never the problem. The image you shipped is.

Follow up on a cold email only when you can add one fact that did not exist when you sent the first one. Not a bump, not a polite reminder that you are still waiting: something the reader did not know before. Two follow-ups written that way beat six written in advance, because a sequence drafted before you send can only ever report that time has passed.

Build articles

An OpenClaw alternative is worth switching to only if it keeps the part you wanted, an agent that acts while you are asleep, and drops the part that should worry you, one always-on assistant holding every credential you own. Three shapes are on offer: another general assistant, a self-hosted workflow runner, or a narrow agent with a scheduled trigger and a single exit door.

Projects

Brands