How long should an email subject line be?

Short enough to survive the phone, which cuts at roughly 35 characters. Desktop clients show around 70 and Outlook around 60, but most mail is now read on a phone, so the phone cutoff is the one that decides whether anybody reads the end of your sentence. Put the thing that earns the open in the first 35 characters and treat the rest as a bonus.

Does starting a subject with Re: increase opens?

It increases opens and it is deceptive under CAN-SPAM, which prohibits a subject line that misrepresents the message. Penalties run to $53,088 per message. A fake Re: also trains the reader to distrust everything else you send, so the gain is one campaign and the cost is the list.

What is a preheader and does it matter?

It is the line the inbox prints next to the subject before anyone opens anything, which makes it the only free copy an email gets. Left empty, the client shows the first line of the message instead, which is usually the greeting or a "view in browser" link. Writing one is the cheapest improvement available to most senders.

Free subject line tester

Two things decide whether a subject works: whether the reader sees the end of it, and whether it trips a rule that stops the message. This shows both.

Subject linePreheader (the line the inbox shows next to it)

This runs entirely in your browser. Nothing you type is sent anywhere, which is why there is no results page to link to.

Where subject lines actually get cut

The number people quote is the desktop one, which is why so many subjects read perfectly on a laptop and land as a half-sentence on a phone.

Gmail on a phone: about 35 characters. This is the number that matters.
Gmail on a desktop: about 70. Fewer and fewer people are here.
Outlook: about 60, and it shows the preheader more prominently than Gmail does.
Apple Mail shows two lines, so the preheader carries real weight and an empty one wastes it on your own greeting.

The patterns that stop a send, not just hurt it

Most subject advice is about open rates. These are different: they are deceptive under CAN-SPAM, so they carry a penalty rather than a cost, and our own send path refuses them.

Re: or Fwd: on a message that is not one. Misrepresenting the message is the specific thing the law names.
A run of exclamation or question marks. Filters score it heavily and it reads as a sales letter from 2003.
Prize, income and guarantee phrases.“You have won”, “risk-free”, “make money fast”. These co-occur with fraud often enough that filters treat them as evidence.
An entirely capitalised subject. Shouting to a person, promotional to a filter.

What no tool can tell you

Whether it will land in the inbox. Placement depends on your sending history, the reader’s own behaviour and signals nobody outside the receiver can see. Anything that prints a confident inbox prediction is guessing and charging you for it. This tells you about your subject line, which is the part you control.

These rules are the same ones Action BeeMail applies before a campaign goes out, so a subject that passes here will not be stopped there. Also free: the spam checker, the CAN-SPAM footer generator and the SPF, DKIM and DMARC checker.