10,636 contacts later: outbound that gets replies

Context

The same seed-stage infrastructure company, on the demand side. Years of accumulated outbound: cold email, LinkedIn ads, Meta experiments, lists on lists. Plenty of activity. No confidence about what any of it produced.

The constraint

Nobody could say which segments replied, which messages worked, or which spend bought actual meetings. The honest answer to “what’s working?” was a shrug. So we stopped sending and started measuring.

What we shipped

A custom persona classifier over 10,636 contacts of lifetime outreach, scoring every segment by real response. Message tests: problem-first framing against cost-first framing. A rebuilt cold-email program with personalization rules instead of templates. And a kill list: every channel that couldn’t show a meeting got its budget removed.

What moved

The analysis found beachhead segments replying at up to 7.1%, more than 3x the 2.01% aggregate. Problem-first ads outdrew cost-first 6x on clicks. The rebuilt cold-email campaign hit 4.08% replies against a 0.25% templated baseline. A targeted LinkedIn program converted 208 MQLs into 25 meetings and 9 demos, the only channel producing meetings that quarter. And we killed over $23.5K of ad spend that had produced zero meetings, including a channel where 90% of leads turned out to be fake.

What this means for you

Most outbound budgets pay for volume because nobody has instrumented response. We measure first, spend second, and killing your worst channel is usually the fastest growth win you’ll get. That discipline is the product.