Email Subject Line Generator
Shows where a subject line is cut off on each client, what a filter notices, and mechanical rewrites of your own words rather than invented copy.
Subject line Introducing the LAST CHANCE to get 30% off your first order! 🎉
characters 62
words 12
first three words Introducing the LAST
Where it gets cut off
iPhone Mail, portrait cut: "Introducing the LAST CHANCE to get…"
Gmail app on Android cut: "Introducing the LAST CHANCE to get 30% off yo…"
Outlook, desktop cut: "Introducing the LAST CHANCE to get 30% off your first order!…"
Gmail, desktop fits, 8 characters spare
which matters because a subject line is glanced at in a list rather than read, so the words after the cut are for people who already opened it
What a filter and a reader notice
trigger words last chance
exclamation marks 1
question marks 0
capitals 23.9% of the letters
emoji 1: 🎉
a number yes
a personalisation token no
where "30% off" appears at character 36, after the mobile cut-off, so most recipients never see it
promotional signals 2.5 between them
reading this reads as promotional before a word of it is read, which is a deliverability question as much as a copy one
Preheader Two days left, and a look at what is new this month
characters 51
repeats the subject no
which is the second line in the list, and the most under-used space in an email
The same words, rearranged
without the opening filler LAST CHANCE to get 30% off your first order! 🎉 (46 characters)
inside the mobile cut-off LAST CHANCE to get 30% off your (31 characters)
without the shouting Introducing the Last Chance to get 30% off your first order (59 characters)
with a personalisation token {{first_name}}, LAST CHANCE to get 30% off your first order! 🎉 (62 characters)
which are mechanical rewrites of your words, not new copy: the judgement about which says the right thing is yours
What a test could measure
list 18,000
reported opens 4,320 at 24%
of which pre-fetched about 1,642 at 38% Apple Mail
opens by a person about 2,678, which is the figure a test is really working with
smallest difference a split test could see 1.25% on an even split of this list
as a relative change 5.2%
which means a moderate difference would be visible on this list, which is large enough for the test to be worth running
The line is 62 characters, so on an iPhone in portrait it reads as
"Introducing the LAST CHANCE to get…". Whatever has to land, has to land
in those first 35 characters.
Open rate stopped being a reliable metric in 2021, when Apple Mail
Privacy Protection began pre-fetching images for a large share of
consumer mail. Those pre-fetches register as opens. Test subject lines
on clicks and replies, and treat any open-rate comparison across time or
audiences as approximate.
The words last chance are the kind that raise a filter's score and a
reader's guard. No single word sends mail to spam on its own, and a line
built from several of them is doing the filter's work for it.
A first-name token is the cheapest thing to test, and the one with the
ugliest failure mode: check what the line reads like when the field is
empty before sending it to a list you did not collect yourself.
The preheader is the second line in the inbox list and the most
under-used space in an email. Leaving it unset means a large number of
recipients see "View this email in your browser" next to the subject.
Curiosity gets the open and costs the trust if the email does not
deliver what the line implied. Measured over a quarter rather than a
send, a line that overpromises trains the audience to stop opening.
A subject line cannot fix an email nobody wants. The strongest predictor
of whether a send performs is whether the list asked for it, which is a
permission question rather than a copy one.
Output is valid and updates as you type.
Fix the highlighted fields to update the output.
Start with the uncomfortable part: the metric everybody optimises subject lines for has been broken since 2021. Apple’s Mail Privacy Protection pre-fetches images for a large share of consumer mail, and a pre-fetch registers as an open whether or not anybody looked. On the example, 38 percent of the list opening in Apple Mail turns 4,320 reported opens into about 2,678 that involved a person. Test subject lines on clicks and replies.
The second part is truncation. A subject line is glanced at in a list, and the list cuts it off: about 35 characters on an iPhone in portrait, a little more in the Gmail app, about 60 on a desktop client. Whatever has to land, has to land in the first 35 characters.
The variants here are mechanical rewrites of your own words, not invented copy. Front-loading a keyword, cutting to the mobile limit, removing the shouting, adding a token. A tool that writes the line for you is writing somebody else’s line.
How to use
- Put in the subject line and the preheader.
- Put in the word that has to land, to see where it falls against the cut-off.
- Add the list size, open rate and Apple Mail share to see what a split test on this list could actually detect.
Example
A promotional line of 62 characters, with a 18,000-person list at 24 percent reported opens:
Where it gets cut off
iPhone Mail, portrait cut: "Introducing the LAST CHANCE to get…"
Gmail app on Android cut: "Introducing the LAST CHANCE to get 30% off yo…"
Gmail, desktop fits, 8 characters spare
What a filter and a reader notice
trigger words last chance
exclamation marks 1
capitals 23.9% of the letters
emoji 1: 🎉
where "30% off" appears at character 36, after the mobile cut-off, so most recipients never see it
promotional signals 2.5 between them
reading this reads as promotional before a word of it is read
The same words, rearranged
without the opening filler LAST CHANCE to get 30% off your first order! 🎉 (46 characters)
inside the mobile cut-off LAST CHANCE to get 30% off your (31 characters)
without the shouting Introducing the Last Chance to get 30% off your first order (59 characters)
What a test could measure
reported opens 4,320 at 24%
of which pre-fetched about 1,642 at 38% Apple Mail
opens by a person about 2,678, which is the figure a test is really working with
smallest difference a split test could see 1.25% on an even split of this list
as a relative change 5.2%The offer sits at character 36, one past the point where a phone stops showing it. That single line is usually the whole finding.
Pitfalls
Open rate is not a clean metric any more. Apple Mail Privacy Protection inflates it by an amount that depends on your audience’s devices, so open-rate comparisons across time or between audiences are approximate at best.
The first 35 characters are the subject line. Everything after them is for people who already opened. Front-load the offer, the name, or whatever the email is actually about.
Spam words do not send mail to spam. Filtering is mostly authentication, sending reputation and engagement. A line stuffed with trigger words does not help, and a clean line does not rescue a domain with broken SPF or DKIM.
Promotional signals add up. One exclamation mark is nothing. An exclamation mark, an emoji, a trigger phrase and a third of the letters in capitals is a promotional tab placement before anybody reads it.
The preheader is the most wasted line in email. Unset, a large number of recipients see “View this email in your browser” next to the subject. Repeating the subject wastes it a different way.
A personalisation token fails badly. Check what the line reads like when the field is empty before sending to a list you did not collect yourself.
Curiosity costs trust over time. A line that overpromises earns one open and trains the audience to skip the next one, which shows up over a quarter rather than in a send.
Compatibility
Runs in the browser: nothing is uploaded and nothing is stored.
Characters are counted as the reader sees them rather than as code units, so an emoji counts as one. The truncation points are approximations of common clients, since the real cut depends on font, device width, orientation and whether a sender name is shown, and they are close enough to make the decision they exist for.
The variants are stated transformations, not generated copy, which makes them testable: the mobile cut never breaks a word, the front-load variant is only offered when the keyword is absent, since removing it from the middle of a line produces something ungrammatical, and the token variant leaves an acronym alone rather than producing “lAST CHANCE”.
The smallest detectable difference is the 95 percent interval on a difference between two proportions at the open rate you gave, on an even split of the list. It is what the test could see, not what it would find.