Feldman, 91,194 emails for Safe Haven Defense, 74 opportunities
Five verticals tested, four cut, and the schools segment carried almost the whole result.
Jay Feldman posted the numbers on a five month cold email campaign for Safe Haven Defense, a company selling retrofit security window film and glass laminate systems that strengthen existing glass so a buyer can skip a full window replacement. From November 2025 to March 2026 the campaign sent 91,194 emails and returned 454 unique replies, 74 tracked opportunities and $370K in tracked opportunity value.
After what Feldman describes as human review of every reply, 78 came back positive, and 52 of those were from people with direct decision-making authority. That is the number worth holding onto. A raw reply count on a 91,194 send campaign tells you very little about whether anyone who can sign is in the thread.
The cut did the work
Feldman says the campaign opened by testing five verticals at once, schools, hospitals, convenience stores, pharmacies and luxury retail, using neutral sender identities such as "School Safety Advisory". Hospitals bounced. The other verticals had low volume. Schools stood out, and inside schools the two titles that converted best were Heads of School and Directors of Facilities.
From there the offer narrowed to one line, a "live security demo near your city", and the whole operation went 100% schools with four spintax variations (spun phrasing variants used to vary each send) and a second campaign running alongside. Feldman says those two school campaigns drove almost all of the result and that the playbook is still working.
What the numbers are and are not
$370K in opportunity value is pipeline, not closed revenue, and Feldman labels it as tracked opportunity value rather than cash. He gives no close rate and no deal count. The 74 tracked opportunities against 91,194 sends works out to roughly one opportunity per 1,200 emails, which is the kind of ratio a practitioner can actually plan capacity against, as long as nobody reads the dollar figure as money in the bank.
The sequencing is the transferable part, and it matches the shape of Feldman's earlier dental campaign writeup. Test several segments on a neutral sender, read the bounce and volume data rather than the copy, and then put everything behind the one segment that answered. Feldman's own summary of the whole thing is blunt about where the lift came from, and it is not from writing.
For people sending, the claim here is not that security window film is a good niche. It is that four of five tested verticals were dropped before the winning campaigns were built, and the result is a product of that cut.
More volume was not the answer. The right audience and one clear offer was.
Know what is landing before you burn another domain
One email a day. Under three minutes. What the operators found out yesterday.