Cold offer architect

Applies before copy, not after. Produces an offer architected as a deal, plus a pass/fail score against the eight-criterion gate below.

Fix in this order, or you'll rewrite copy forever

Market → offer → copy → list → volume. When something isn't working, work down that list and stop at the first thing that's broken. Almost every "our cold email isn't working" problem is an offer problem being treated as a copy problem, which is why the rewrite never helps.

On cold traffic the offer moves more than copy, personalization, and volume combined. A strong offer with plain copy beats a weak offer with brilliant copy, every time.

An offer is not an opener

"Worth a quick call?" · "Can I send it over?" · "Open to a chat?" — those ask for interest in exchange for nothing the reader can value yet. They are openers.

An offer is a complete value-for-value deal: a specific outcome they already want, a mechanism that makes it believable, the risk moved off them, and the terms. Architect the deal first. The message is only the vehicle that delivers it. Write the email before the deal exists and you will produce a polite opener every single time.

The gate — eight binary criteria, all must pass

Any "no" means not shippable. Score honestly; do not average.

#	Criterion	A "no" looks like
1	Specific, quantified, time-bound outcome	"more pipeline", "better data" — no number, no clock
2	Believable mechanism	A skeptic can't see why this works, or why you
3	Risk sits on the seller	The buyer spends money, time, or face to find out
4	Near-zero effort for the prospect	A meeting before any value lands — that meeting is the effort
5	A reason why it's this good	Generosity with no explanation reads as a trick
6	Proof embedded, not deferred	"Happy to send a case study" — deferring the one thing that earns belief
7	Leads with them	The first line contains "we are / we do / we help"
8	Uncomfortable to make	If you're perfectly comfortable offering it, it's too small

A soft "want it?" opener typically fails 1, 3, and 8 at once. That's why it feels weak even when the writing is clean.

When a criterion fails, pull its lever and re-score — the ladders for risk, proof, effort, and specificity are in references/amplification-ladders.md. To find which criterion is causing a live campaign to die, work the symptom table in references/diagnose-a-dead-offer.md.

The forcing loop

Architect the deal → score all eight → fix the lowest one → re-score. Repeat until every criterion passes. Never present an offer that hasn't cleared the gate; a 6/8 shipped is a month of sending burned to learn something the gate already told you.

What good looks like

The tell of a good operator: they can name which of the eight is currently weakest without being asked, and they treat offers like ad creative — many, tested, killed fast — rather than as a brand decision made once.

The mediocre version is a well-written email around a hollow deal: correct structure, real personalization, a polite ask, and a value proposition the reader has no reason to believe or care about. It gets praised internally and produces nothing.

Good output makes a fitting stranger think one of: "this is a no-brainer", "I'd be an idiot to pass", or "what's the catch?" — and the catch is answered inside the offer. If your ideal prospect could comfortably ignore it, it isn't done.

Rules

• MUST architect the deal before writing any message.
• MUST embed a specific proof point inside the offer — a real number, and a named comparable where naming is permitted.
• MUST deliver value before asking for a meeting, never the reverse.
• NEVER present an offer that fails any of the eight criteria.
• NEVER promise an outcome you can't defend, and never invent a number to pass criterion 1.