The 7 Point Cold Email Rubric We Score Every Sequence Against
Most cold email feedback is taste, someone senior says "make it punchier" and the BDR guesses. Here's the fixed seven-point rubric we score every sequence against instead, with a failing and passing example for each.
Aditya Dubey
September 24, 2026
8 min read
Most cold email feedback is taste. Someone senior reads a draft, says "make it punchier," and the BDR guesses at what that means. Two weeks later the same note comes back on a different email, because nobody wrote down what punchier was.
We score written sequences against seven criteria, the same seven every time. A BDR does not go live until a sequence clears the bar. The value is not that the criteria are clever, most of them are obvious. The value is that they are fixed, so a BDR's week one copy is comparable to their week four copy, and a note means the same thing every time it is given.
Here they are, with an example of a draft failing each one and what it becomes after the rewrite. The examples are written to illustrate the criterion rather than taken from client sequences.
1.
Subject line Text:
Specific, under five words, lowercase. Vague loses regardless of formatting.
2.
Personalisation Text:
The detail has to change what you say next, not just prove you looked.
3.
Word count Text:
Under 80 words per touch. The whole email visible without scrolling.
4.
Proof point Text:
One checkable customer detail beats any number of adjectives.
5.
Conversational tone Text:
Read it aloud. If you would not say it across a table, cut it.
6.
Email flow Text:
One ask, with an easy honest exit. Three asks is no ask.
Once these six checks are in place, the next step is to make sure every touch works together as a sequence.
7.
Sequence logic Text:
Each touch is a different angle on the same problem, never a bump.
1. Subject line Fails: Quick question -
The most common subject line in B2B outbound and the least informative. It promises nothing, so it competes on curiosity alone against forty other emails doing the same thing. Worse, it sets up an email that has to justify the mystery.
Before (Fails):
Quick question
vs
After (Passes):
After-hours enquiries at [Company]
Three rules, in order of how much they matter.
Name the thing the email is about, in the reader's vocabulary. Not clever, not vague. If the reader knows what the email concerns before opening it, the ones who open it are the ones who care.
Keep it under five words. Mobile clients truncate somewhere around 35 characters, so anything longer is being judged on its first half anyway. Write the first half.
Lowercase it. A capitalised, title-case subject line looks like every marketing email in the inbox, because marketing emails are written in a CMS by someone who capitalises headlines. Lowercase reads like a colleague typing quickly. It is a small pattern break and it costs nothing, but it only works while it is uncommon, so treat it as a current tactic rather than a law.
One thing lowercase does not fix: a vague subject line in lowercase is still vague. The specificity rule comes first and the other two are finishing.
2. Personalisation -
Before (Fails):
Hi Sarah, loved your recent post on AI in healthcare. Really insightful stuff. On a related note, we help healthcare companies automate patient communication.
vs
After (Passes):
Hi Sarah, your post on AI in healthcare made the point that most automation stops at scheduling. That is usually where it stops for the teams we work with too, and the reason is the compliance review on anything that touches clinical language.
This is proof that the BDR looked. It is not proof that the BDR thought. The compliment and the pitch are stapled together by "on a related note," which is the seam showing. Every reader has received this exact email.
The observation has to lead somewhere. If you can delete the personalised sentence and the email still makes sense, it was decoration. The test: does the thing you noticed change what you are about to say?
3. Word count -
Before (Fails):
143 words explaining who you are, what the product does, three benefits, and a meeting request.
vs
After (Passes):
Under 80 words per touch.
The constraint is not aesthetic. A cold email is read on a phone, in a gap, by someone who did not ask for it. Under 80 words, the entire email is visible without scrolling and the ask is unmissable. Over 80, the reader makes a decision based on the shape of the block rather than its content.
Cutting to 80 also forces the harder edit. You cannot fit three benefits, so you have to know which one matters to this person. Most padding is the writer avoiding that choice.
4. Proof point -
Before (Fails):
We're trusted by leading healthcare brands to transform patient engagement at scale.
vs
After (Passes):
One customer, a 40-clinic group in Texas, was routing after-hours enquiries to a voicemail box that someone cleared at 9am. They now get a response inside four minutes.
Nothing in the failing version can be checked, so nothing in it is believed. "Leading," "transform," and "at scale" are each doing the job of hiding the absence of a number.
Small and specific beats large and vague. A checkable claim about one customer is worth more than an unfalsifiable claim about many. If the only proof available is genuinely vague, leave the proof point out entirely rather than filling the slot with adjectives.
5. Conversational tone -
Before (Fails):
I wanted to reach out to explore whether there might be an opportunity to leverage our solution to drive efficiencies across your patient communication workflows.
vs
After (Passes):
Most clinics we talk to handle after-hours enquiries the next morning, and nobody loves it but it works. Curious whether that is where you have landed too.
Nobody speaks like the failing version. Register that formal reads as a template, because templates are the only place it appears.
The test is mechanical: read it out loud. If you would not say the sentence to the person across a table, it does not go in the email. "Leverage," "solutions," "reach out," and "synergies" almost never survive that test.
6. Email flow -
Before (Fails):
Happy to send over a case study, or we could do a quick 15 minutes this week, or if it's easier I can just share our one-pager. Let me know what works!
vs
After (Passes):
Worth a short call, or is this not a priority right now?
Three asks is no ask. The reader has been handed a decision with three branches at the exact moment you wanted them to do one easy thing, and the easiest branch is always to close the email.
One ask, and one that gives the reader an easy honest exit. Giving people permission to say no raises reply rates, because a no is also a reply and tells you to stop.
7. Sequence logic -
Before (Fails):
Touch three reads "just bumping this up in case you missed it."
vs
After (Passes):
Each touch carries a different angle on the same problem. The first opens with the observation. The second brings the proof point. The third changes the question rather than repeating it, for example asking whether this sits with them at all or with someone else. The fourth closes the loop and stops.
Bumping is not a new reason to reply. It repeats the original ask with added guilt, and it tells the reader the sender has nothing further to say.
Sequence logic is the criterion BDRs skip most, because it cannot be judged from a single email. It is the reason we review sequences as a unit rather than touch by touch.
💡The two that usually fail together
Word count and proof point tend to fail in the same draft, and for the same reason. A BDR who is not confident the email is credible adds sentences to shore it up, and the sentences they add are the vague ones, because specific proof is harder to find than adjectives. The email gets longer and less believable at the same time.
The fix is almost never "write shorter." It is finding one real, checkable detail. The length problem solves itself once the credibility problem does.
Where a rubric does not help
A sequence that clears all seven criteria and still gets no replies is not a copy problem. It is usually a list problem, and rewriting the email a fourth time is the most comfortable way to avoid finding that out.
The other possibility is that nobody is reading it because it never arrived. Sending volume per mailbox, domain warm-up, authentication records, and inbox rotation decide whether well-written copy reaches a primary inbox at all, and none of it shows up in a rubric score. That is a separate discipline and a separate post.
The rubric also will not tell you whether a sequence is right for your buyer, only whether it is well built. A perfectly scored email to the wrong persona is a perfectly scored email that fails.
And the criteria only work if the same person, or one aligned group, applies them. Split scoring across several reviewers with slightly different standards and week one stops being comparable to week four, which was the entire point.
Get your sequence scored
If you want your current sequence scored against these seven, send it over and we will mark it up and send it back, including the criteria it passes.
Aditya Dubey