Cold email subject line tester

Most subject line testers in search results promise more opens, which is how newsletters and promotions are judged. A cold email asks one person for a reply, and an open has become weak evidence: Apple says Mail Privacy Protection prevents senders from seeing whether a recipient opened their message.

ReplyLead does not collect the signal at all. Open tracking was disabled on all 115 campaigns with sends in its 12 August 2026 benchmark extraction, so there is no ReplyLead open rate to calibrate a score against, and this page does not borrow one from elsewhere.

What ReplyLead measures is replies: a 2.12% median of unique replies per contacted lead among 81 campaigns with at least 500 contacts, extracted 12 August 2026, automatic replies included. That figure describes whole campaigns, not subject lines; the reply rate benchmark shows how it was calculated.

The Internet Message Format standard describes the Subject field as a short string identifying the topic of the message. The subject tests hold a line to that job: name the topic and the reader, or a concrete outcome, in words a colleague could have typed.

Spintax rotates wording so that one campaign does not send the same string to everyone. Each option still goes to a share of the list, so a weak option is a weak first impression for those people. It does not average out against the strong ones. The tester expands every option instead of picking one at random, so each version is checked before it sends.

It reads the common forms: {a|b} groups, nested groups, empty options such as {{{first_name}}, |}, the {{RANDOM | a | b}} form some platforms use, double-brace fields such as {{company}} with an optional Liquid default, and single-brace fields such as {COMPANY}. ReplyLead's spintax guide covers the syntax and why variation is not a substitute for a better line.

Inbox lists cut a long subject with an ellipsis, and where the cut falls depends on the device, the app and the font. The tester marks the point where about 300 pixels of 15-pixel system text end, roughly the subject width on a phone about 390 pixels wide, and dims what follows. The width is measured in the browser you are using, in the same font the inbox rows use.

A cut is shown, not scored. A subject whose first words carry the topic still works when the end is hidden. In the inbox, the preview text beside the subject comes from the first sentence of a plain-text email; email preview text explains how inboxes build it.

The subject tests, the 247 template openers and the em dash rule come from the script of ReplyLead's cold email grader. The 1,130 phrases with their weights, the ALL CAPS rule and the exclamation and length rules come from the spam checker. Both were read from the live pages on 3 November 2026, and the build stops if either page's rule code no longer matches.

The tester applies the subject tests to each version as written, with merge fields still in place, the way the grader reads a subject. Every other rule runs on the text a sample lead would receive. The spam checker treats its 246 weight-1 phrases as mild everyday words that matter only in combination, so here they mark a version Check rather than Fails.

Expansion stops at 200 versions and the inbox shows the first 50; the sample-lead totals count the same 200. Everything runs in your browser: the subject you type is not sent to ReplyLead or stored, and analytics records only counts, such as how many versions passed.

These are editorial checks. They do not test inbox placement, read your sending domain or predict replies.