Cold email subject line grader, scored on 901,002 prospects' replies
By Ryan Doyle, co-founder of Sales.co. Published October 23, 2026, updated October 23, 2026.
Put the prospect's first name in the subject line. Across 901,002 prospects Sales.co emailed from Oct 6, 2025 to Oct 8, 2026, lines with the first name got 0.88% human replies against 0.56% without it, 1.6 times as many. Five to seven words did best (0.79%), one or two words did worst (0.50%), and all-lowercase lines trailed (0.53%). Lines with a question mark, a number or the company's name showed no consistent difference.
For B2B founders and sales teams who write their own cold email: type a subject line, and the grader checks it against the patterns that went with more human replies in that data, then says what to change.
What this page adds: a grade built on replies from one company's own sending: 901,002 prospects and 6,116 human replies. Each subject line was read once and turned into the patterns below; none is published.
Every pattern and its reply rate
One row per pattern value. A prospect counts once: the subject line is the one on the first email they got from that sender, and a reply counts if any human reply came back from their address.
| Pattern | Value | Prospects | Human reply rate | Interested rate | Human rate, largest sender removed |
|---|---|---|---|---|---|
| Prospect's first name in the line | Yes | 340,105 | 0.88% | 0.30% | 0.77% |
| No | 519,495 | 0.56% | 0.18% | 0.54% | |
| No first name in the address | 41,402 | 0.57% | 0.23% | 0.55% | |
| Word count | 1 to 2 words | 78,318 | 0.50% | 0.17% | 0.50% |
| 3 to 4 words | 779,391 | 0.69% | 0.23% | 0.63% | |
| 5 to 7 words | 35,830 | 0.79% | 0.39% | 0.76% | |
| 8 or more words | 7,463 | 0.54% | 0.32% | 0.50% | |
| Length | 1 to 15 characters | 102,300 | 0.70% | 0.24% | 0.65% |
| 16 to 30 characters | 751,172 | 0.68% | 0.23% | 0.63% | |
| 31 to 45 characters | 40,307 | 0.59% | 0.27% | 0.57% | |
| 46 to 60 characters | 5,225 | 0.54% | 0.42% | 0.56% | |
| 61 or more characters | 1,998 | not shown | not shown | not shown | |
| Capitalization | First letter capital only | 476,072 | 0.71% | 0.21% | 0.60% |
| Title Case or mixed | 383,548 | 0.66% | 0.26% | 0.68% | |
| all lowercase | 40,905 | 0.53% | 0.13% | 0.52% | |
| No letters | 477 | not shown | not shown | not shown | |
| Question mark | Yes | 63,256 | 0.71% | 0.27% | 0.59% |
| No | 837,746 | 0.68% | 0.23% | 0.63% | |
| Contains a number | Yes | 26,231 | 0.66% | 0.28% | 0.65% |
| No | 874,771 | 0.68% | 0.23% | 0.63% | |
| Prospect's company in the line | Yes | 170,857 | 0.64% | 0.24% | 0.62% |
| No | 725,577 | 0.68% | 0.23% | 0.63% | |
| No company in the address (free mail) | 4,568 | 1.12% | 0.44% | 0.69% | |
| Starts with Re: or Fwd: | Yes | 19,900 | 2.02% | 1.41% | 0.72% |
| No | 881,102 | 0.65% | 0.20% | 0.63% |
All prospects: 0.68% human replies, 0.23% interested (largest sender removed: 0.63%). "Not shown": fewer than 2,000 prospects, or one sending workspace made more than 45% of them. Source: Sales.co sending data, October 6, 2025 to October 8, 2026.
What the grade checks
The grade counts the patterns that went with more replies both in all the data and with the largest sender removed:
- Add the prospect's first name: 0.88% human replies against 0.56% (largest sender removed: 0.77% vs 0.54%).
- Use 5 to 7 words: 0.79% human replies against 0.67% (largest sender removed: 0.76% vs 0.62%).
- Keep it to 30 characters or fewer: 0.68% human replies against 0.59% (largest sender removed: 0.63% vs 0.57%).
- Capitalize the first letter (not all lowercase): 0.68% human replies against 0.53% (largest sender removed: 0.63% vs 0.52%).
Meeting all of them is an A, three in four a B, and half or two in three a C; fewer earns a D grade. The first-name check runs only when you type the prospect's first name, the way the data could check it only when the address held one.
Limits of this data
The strongest objection: a subject line travels with its list, its offer and its sender. A team that writes the first name into every line may also pick tighter lists. These are associations across 59 senders, and no line was split-tested against another, so a better grade is a reason to test a line on your own list before you roll it out.
What to do with this
Write five to seven words, keep them under about 30 characters, capitalize the first letter and put the prospect's first name in. Then spend the rest of the effort on the list and the first line of the email. For examples to adapt, see 80 cold email subject lines; for when to send them, the best time to send cold email; for what comes after, the follow-up email templates.
If you would rather have someone write and send it, Sales.co runs done-for-you cold email from $2,000 a month for up to 10,000 emails, month-to-month. See pricing.
Method
- Source: Sales.co's sending database, one record per cold email sent, and the replies the sending tool labeled. Unit: recipient thread: one recipient address in one sending workspace; its subject is the subject of the first email that address got; 901,002 of them, October 6, 2025 to October 8, 2026.
- A human reply means the prospect replied at least once with a reply that is not an auto-reply, out-of-office, bounce, warm-up, test or spam (sending tool's labels). An interested reply means at least one of those replies was labeled 'possibly interested'.
- First name: the subject contains the first name taken from the recipient's address (john.smith@ -> john); 'unknown' when the address has no usable first name. Company: the subject contains the company taken from the recipient's domain (acme.com -> acme); 'unknown' for free mail. Length counts characters after any Re:/Fwd: prefix.
- A reply counts only when it came from the address that was emailed, so replies from a colleague or another inbox are missed and every rate here is a floor. The overall rate (0.68%) is per prospect, so it runs higher than the per-email rate on the send-time page, where a prospect who got three emails counts three times.
- Left out: any pattern value with fewer than 2,000 prospects, or where one sending workspace made more than 45% of them. The largest sender made 25% of all prospects, so the table also shows each rate with it removed. No client is named and no subject line, email or address is published.
- Refresh: each time the data is recomputed, this page is regenerated from it; the freshness line at the top is written from the data file.
Cite this: Sales.co, "Cold email subject line grader", https://sales.co/cold-email-subject-line-grader, data through October 8, 2026.
Where to go next
Cold email subject lines: 80 templates · best time to send cold email · cold email statistics · lead generation calculator · all Sales.co research
Sources
- Sales.co sending database: 901,002 prospects and 6,116 human replies, October 6, 2025 to October 8, 2026, as of October 23, 2026.
- Sales.co pricing and plans: sales.co/pricing, read October 23, 2026.
FAQ
What makes a good cold email subject line?
In Sales.co's data, four things went with more replies. The prospect's first name (0.88% vs 0.56%). Five to seven words (0.79% vs 0.69% for three or four). 30 characters or fewer (0.68% vs 0.59%). A capital first letter rather than all lowercase (0.71% vs 0.53%). Each held with the largest sender removed.
How long should a cold email subject line be?
Five to seven words and no more than about 30 characters. Lines of 5 to 7 words got 0.79% human replies; 1 to 2 words got 0.50% and 8 or more got 0.54%. Lines of 30 characters or fewer got 0.68% against 0.59% for longer ones.
Should you put the prospect's name in a cold email subject line?
Yes. Lines with the first name got 0.88% human replies and 0.30% interested replies, against 0.56% and 0.18% without it. With the largest sender removed the gap holds: 0.77% vs 0.54%. 40% of the lines where a first name could be checked already used it.
Do questions in cold email subject lines get more replies?
Not in this data. Lines with a question mark got 0.71% against 0.68% without, and with the largest sender removed the order flips (0.59% vs 0.63%).
Do lowercase subject lines work for cold email?
They did worst of the capitalization styles here: 0.53% human replies for all-lowercase lines, against 0.71% when only the first letter was capitalized.
Does a fake Re: in the subject line help?
It looks like it in the raw numbers (2.02% vs 0.65%), but the largest sender made 44% of those lines, and without that sender the gap shrinks to 0.72% vs 0.63%. A Re: on a first email also tells the reader something false, so this grader does not reward it.