The machine, in public
Paste the email you are about to send. The rule-set we trained on 4,000,000 sends scores it and a full graded report lands in your inbox in minutes: what would get it deleted, what to keep, and your email rewritten in full. A few hours later a full three-step sequence, written for your ICP, follows from Hugo. Free.
Questions
The same mechanical rules our own copy passes before a sequence ships: sentence count against the five-sentence ceiling, body length against the under-40-words band, subject line length and clichés, researched merge fields against the two-to-six plateau, reading grade against a 6.0 ceiling, machine-writing tells like the em dash, and the phrases that read as template, such as synergies, quick question, hope this finds you well, and their relatives.
Your sequence is scored in memory, read once by the engine to write your report, and read by one person for the human version. It is never added to any corpus and never used for training. Your address is used to send the report and so a real person can follow up.
From more than 4,000,000 cold emails we sent over three years and what came back. Two findings carry measured effect sizes: emails over five sentences convert about 30% worse (measured across 32 client accounts), and first touches under 40 words converted at roughly double the rate of 40-to-69-word emails (across 44 clients). The rest is mechanical hygiene, labelled as such.
No, and anyone who tells you otherwise is selling something. Copy explains a small share of conversion variance; the offer, the list and the deliverability carry the rest. The score tells you what would get the email deleted before it was read. It cannot tell you your offer is good.
Yes. The subject line is scored separately: word count against the three-to-five-word band, all-caps and fake-Re: detection, and the subject clichés every prospect has already deleted this week.
EMPRA