How to score your cold email in 2026: a 6-dimension rubric you can run on any draft

If you have ever stared at a cold email draft wondering "is this any good?", the problem is rarely the words. It is that you have no rubric. Without one, you tune toward whatever caught your eye in the last newsletter you read, which is usually subject-line tricks or AI personalization fads. The result is a draft that scores well on the thing you obsessed over and badly on everything else.

Here is a 6-dimension rubric we use on every Coldsmith outbound, plus how to score your own draft right now without installing anything.

The 6 dimensions, ranked by impact

The rubric weights personalization 2x and ask specificity 1.5x because those two account for the bulk of reply-rate variance in the 1,000+ cold emails we have audited. Length, tone, value clarity, and send-readiness each get 1x weight; below a certain floor they tank reply rate (a 400-word email is dead on arrival), but above that floor the marginal return drops fast.

1. Personalization (weight 2x)

The question: does the first line cite a specific, recent signal about the recipient?

Scoring:

• 10: The opener cites a dated, verifiable detail (a post, a launch, a quote, a hire) the prospect would recognize. Sender clearly did the homework.
• 7: A signal is cited but stays at the topic level ("I saw your team is hiring") rather than something specific.
• 4: Compliment-led opener ("love what you are building") with no concrete reference.
• 0: Mail-merge template ("I came across your impressive work at {company}").

The leverage here is enormous. A 10/10 personalization line lifts reply rate 3 to 5x over a 4/10 line on the same email body. It is the single biggest determinant of whether the prospect reads sentence two.

2. Ask specificity (weight 1.5x)

The question: is the CTA concrete, small, and time-bound?

Scoring:

• 10: "15-min call next Tuesday at 11am ET" or "reply yes/no, no call needed". Specific time, low commitment.
• 7: "15-min call next week" or "Open to a quick chat?". Concrete enough but no time anchor.
• 4: "Would love to connect" or "Let me know your thoughts".
• 0: No CTA, or three different asks in the same email.

A high specificity score forces the prospect into a binary decision instead of a vague maybe. Most cold emails drop here even when the opener is great, because the writer hedges at the close to avoid sounding pushy.

3. Length

The question: is it under 120 words, scannable on mobile?

Scoring:

• 10: Under 80 words, three short paragraphs max.
• 7: 80 to 120 words.
• 4: 120 to 200 words.
• 0: 200+ words or any single paragraph longer than 4 lines.

Cold emails compete with replies and Slack notifications on a phone screen. Anything that does not fit in the preview pane gets archived.

4. Tone match

The question: does it sound like a peer reaching out, not a marketing template?

Scoring:

• 10: Reads like a founder DM'ing another founder.
• 7: Mostly natural with one or two template tells ("Hope this finds you well", "Just wanted to reach out").
• 4: Multiple template tells; clearly cold-email-template-flavored.
• 0: Reads like an SDR sequence ("I help X companies do Y", "We work with Z").

The tone-match score is the strongest predictor of whether the recipient pattern-matches your email as "real human" vs "sequence." Once they decide it is a sequence, they archive on instinct.

5. Value clarity

The question: is what you are offering clear in one sentence?

Scoring:

• 10: The recipient knows exactly what you do and why it matters to them, in one read.
• 7: Clear, but the "why it matters to them" requires inference.
• 4: Vague offer ("we help with growth").
• 0: No clear value mentioned, or three different value props.

A clear value prop respects the reader's time. A vague one signals you have not thought through who they are.

6. Send-readiness

The question: any deliverability or trust red flags?

Scoring:

• 10: No links, no images, no attachments. Plain text style with a real reply-to address.
• 7: One link to a relevant page; no attachments.
• 4: Multiple links, or one image, or salesy phrases ("limited time", "act now").
• 0: HTML-heavy template with tracking pixels, multiple links, or attachments.

Email providers and recipients both flag the same patterns. A well-written email with tracking pixels and three CTAs in a styled HTML template will land in promotions tab or worse, even if the content is great.

Score it yourself in 30 seconds

You can run this rubric manually in about 5 minutes per email, which is fine for one or two drafts. Past that, paste your draft into coldsmith.dev/tools/cold-email-audit and the same rubric returns a 0-100 score plus per-dimension verdicts and the single biggest lever to pull. Free, no signup. You get a shareable URL for the result, in case you want to compare scores with a teammate.

The audit is the same code we run on our own outbound before sending. The 6-dimension rubric is the same one we use to grade the leads we deliver in our $49 list orders, where every row's first-liner has to score 7+ on personalization or it does not ship.

What the median cold email scores

Across 1,000+ audits in our database, the median cold email scores 53 out of 100. Distribution:

Most cold emails are in the C/D range. The gap between a 55 and a 75 is almost always personalization and ask specificity. If you fix those two, the email moves from "I will archive this" to "okay, sure, 15 minutes" in the recipient's head.

What a 75+ email looks like

A real example, redacted (cited a specific signal, asked for one small thing, no template tells):

Hi M, noticed your team shipped the edge KV product last week, the "100ms global writes" framing is sharp positioning against the legacy KV competition. We dropped Brex's email infra bill 38% in one quarter by moving transactional sends off SendGrid onto a worker-based dispatcher; same pattern could apply to your new launch's transactional volume. Open to a 15-min call next Tuesday at 11am ET to walk through the migration?

Personalization 9, specificity 9, length 9 (under 80 words), tone 8 (one minor template echo in "same pattern could apply"), value clarity 9, send-readiness 10. Total: 88/100, grade A.

What a 45 email looks like

Hi {first_name}, hope you're doing well! I came across your work at {company} and was impressed. I help SaaS companies like yours grow their pipeline through targeted outreach. Would love to connect for a quick call to discuss how we can help. Looking forward to hearing from you!

Personalization 1, specificity 3 (vague CTA), length 8, tone 2 (textbook SDR template), value clarity 3, send-readiness 10. Total: 39/100, grade F.

The difference is not effort, it is process. The 88-score email took 5 extra minutes of research. That 5 minutes is what we charge for in our $49 list orders, applied across 100 prospects per ICP.

Where to go next

If you write your own cold emails: bookmark the rubric above and audit your last few drafts. Most teams find one or two dimensions consistently dragging them down. Fix those first; do not optimize the high-scoring ones further.

If you want a researched list where personalization is already done: grab a free sample pack to see what a 9/10 first-liner looks like in our format, then order a 100-lead starter at $49. 24h turnaround, 7-day refund if a meaningful share of rows misses the quality bar.

If you want the writing patterns themselves: the First-Liner Playbook ($9) has 50+ opener templates organized by niche, each one annotated with which dimensions it nails and which it sacrifices.