The archive only shows failures because failures get submitted. This section corrects the imbalance - annotating high-scoring messages and the specific signals that earned them.
A message that opens with a verifiable signal, names a specific ICP, leads with a real customer outcome, and asks for something proportionate to what it established. One minor offense - a premature calendar ask - prevents an A grade.
Opens with strong specific research, names a precise ICP. Falls short on the value prop - the outcome claim exists but lacks a number.
Four things. A specific, verifiable reference to the recipient. An ICP narrow enough that most companies don't qualify. A real customer outcome with a real number. And an ask that matches the value shown, not the sale being chased.
The gap between a D and a B is almost never the product. It's the structure. Same product, same recipient, but with actual research and a right-sized ask, the score jumps 4-5 points.
The best cold email we've seen scored a B (6.9/10). Nobody has hit an A. Most common sticking point: even strong messages include a calendar request that's bigger than what the message earned.
Messages that name a specific customer with a specific number score 2.3 points higher on average than messages with vague social proof. One real number beats three logo drops.
Four things. It mentions something specific and verifiable about the recipient. It names an ICP narrow enough to exclude most companies. It includes a real customer outcome with a real number. And the ask is sized to match the value shown, not the sale the sender wants.
The best we've seen is a B (6.9/10). Nobody has scored an A yet. The most common reason: even strong messages include a premature calendar request that doesn't match the value delivered in the body.
Almost never the product. It's the structure. Same product, same recipient, but with real research and a right-sized ask, the score jumps 4-5 points. One named customer with a number beats three logo drops every time.