COLD OUTBOUND TEARDOWN · DIAGNOSIS ONLY, NO REBUILD
Source reviewed: Two cold outbound drafts, LinkedIn DM and email, shared for review
Scored on all 7 points. Neither clears half.
By Alex Iliescu — Strategic Flow · structural read only, no fix included
4/10LinkedIn DM score
5/10Email score
0Calls requested, either draft
1/2Drafts with a yes-path
How this is read, and what's scored
Two separate drafts, scored independently, then compared. Each is read once as a cold recipient would read it, checked against four structural points. No rewrite is included. This is diagnosis only.
All 7 points of the Decision Friction Model, run against both drafts. No point skipped, no partial framework. Diagnosis only, no rebuild, that's a separate deliverable.
01Subject lineFiling label or consequence signal? A DM has no subject field, so the opening line is scored as its structural equivalent.
02Lead constructionConsequence before caveat, or context before consequence?
03Feature-to-outcome translationDoes it name what changes for the reader, or what the product does?
04Visual hierarchyDoes the most important claim lead, or is it buried, or is there no hierarchy at all?
05Proof placementIs the strongest stat, quote, or named result above the fold, or absent entirely?
06Social proofIs there a third-party voice, a number, a named result?
07CTA languageOwnership language ("reply run") or guest language, or no true CTA at all?
Failure pattern vocabulary
Every fail below is one of these exact patterns, no paraphrasing. Only the five that actually appear in these two drafts are listed here.
→Feature-First BiasThe message leads with what the product does, not what the reader gains. Triggered when point 03 fails: mechanism described, no outcome named.
→Consequence-After-CaveatThe outcome is buried behind context, disclaimers, or qualifications. Triggered when point 02 fails: the message concedes ground or sets context before it states its claim.
→Missing HierarchyNo structural priority. Everything looks equally important. Triggered when point 04 fails: no line, claim, or word is weighted above the others.
→Zero Social ProofNo third-party voice, no named result, no specific number. Triggered when points 05 and 06 fail together: nothing in the message proves the claim has ever worked, only that it's being made.
→Guest Language CTAThe CTA describes what the brand offers, not what the reader does, or no true CTA exists at all. Triggered when point 07 fails: the reader has no owned action, only an offer or an opt-out.
LinkedIn DMFull text below
4/10
Two points hold, one is partial. Zero Social Proof and Guest Language CTA are what drag it below half.
"hey {first}, are you writing outbound in claude yourself, or still living in a sequencer?
if it's claude, the piece we ship is the list + who-got-which-step + you approve before linkedin/gmail leaves. not another inbox farm.
if that split is wrong for how you run it, say so and i'll drop it."
✓01. Subject line (opening line, no subject field): "are you writing outbound in claude yourself, or still living in a sequencer?" names the reader's own practice, not a filing label
✓02. Lead construction: opens on the reader's situation before any caveat or context about the product
✗03. Feature-to-outcome translation — Feature-First Bias: "the list + who-got-which-step + you approve" describes what the product does, never states what changes for the reader
~04. Visual hierarchy: no bold or emphasis, but each idea gets its own paragraph, the question, the mechanism, the opt-out are separated, not merged into one block
✗05. Proof placement — Zero Social Proof: no stat, no quote, no named result anywhere in the message
✗06. Social proof — Zero Social Proof: same gap as above, no third-party voice at any point
✗07. CTA language — Guest Language CTA: "say so and i'll drop it" is an opt-out, not an action the reader owns. No CTA exists for a reader who agrees
EmailFull text below
5/10
Best CTA of the two, but Consequence-After-Caveat and Zero Social Proof still hold it right at half.
"subject: your sequencer is fine. the loop isn't
you already have instantly/smartlead. we're not trying to beat that on volume.
what sits in claude today is copy. what should sit outside it is the crm list, stale-title check, and approve-before-send.
if you want that walked on your own list, reply "run" and i'll do the first 10 in this thread."
✓01. Subject line: "your sequencer is fine. the loop isn't" is a consequence signal, not a filing label, it makes a claim, not an announcement
✗02. Lead construction — Consequence-After-Caveat: opens by conceding the reader's existing tool ("you already have instantly/smartlead") before stating what's actually different
✗03. Feature-to-outcome translation — Feature-First Bias: "the crm list, stale-title check, and approve-before-send" is a feature list, no stated reader outcome
~04. Visual hierarchy: no bold or emphasis, but the subject carries the claim and the body is paragraph-separated by idea, same partial structure as the DM
✗05. Proof placement — Zero Social Proof: no stat, no quote, no named result
✗06. Social proof — Zero Social Proof: "the first 10 in this thread" is a promise, not a result already shown
✓07. CTA language: "reply run" is genuine ownership language, one word, the reader initiates it
ComparisonSame sender, same offer, one point gap
Same 7 points, same three failures, different order.
Both drafts fail Zero Social Proof outright, on both of the two proof-related points, and both get partial credit on hierarchy for paragraph separation, plain text has no bold to withhold. That's the real shared gap. The remaining difference is where each one spends its other points.
LinkedIn DM
4/10 · 2 pass, 1 partial
Wins on subject and lead construction, the opening question is genuinely strong. Loses everything downstream of it, including CTA language, where "say so and i'll drop it" gives no reader a way to say yes.
Email
5/10 · 2 pass, 1 partial
Wins on subject and CTA language, the strongest single line across both drafts is "reply run." Loses lead construction by conceding to the competitor before stating the claim.
Both drafts fail the same two points in the same way: Missing Hierarchy and Zero Social Proof. That's not a copy problem in either draft, it's the same structural gap showing up twice. In a benchmark of 67 published emails, zero or buried social proof is the single most common failure pattern, this pair is consistent with that baseline, not an exception to it.
Contra-pitchNot a rebuild, a mirror
This is the diagnostic layer.
The findings above identify where each draft holds and where it loses the reader. The rebuild is the next step, CTA construction, tension resolution, and message-by-message sequencing.
Quick self-check · your own sequence
Check what applies — the count updates live, no need to submit anything.
This is the same engine, not a sample of it. Fixed diagnostic checklist. 67 published audits. Average score moves from 3.4 to 9.0 once the structure gets rebuilt, not just diagnosed. If a sequence you're already sending would survive this same read, that's worth knowing before a reader finds out for you: Decision Friction Review ↗