Every cold email operator has a story about the morning the sending stopped. You open the dashboard, the deliverability graph has fallen off a cliff, and somewhere in a Google Workspace admin console a policy you never read decided your entire operation was a pattern.
Taylor Haren posted his version of that morning this month, and it landed harder than most because he attached a number. He says the loss came to $31,247 of cold email infrastructure. He titled the post around the end of the loophole era. Then he described rebuilding, and said the rebuilt system now sends 300,000 emails a month at a 34% positive reply rate.
Two claims, one post. The first one I believe. The second one I want to look at properly, because those two numbers cannot both be casual.
The post that started it
Set the framing aside and look at the mechanic being described. An agency accumulates sending capacity: domains, mailboxes, warmed IP reputation, a reseller relationship that makes the per-mailbox cost work. That capacity is the product. Losing it is not losing a tool, it is losing the factory.
None of that is verifiable from the outside. It is a practitioner claim on a social platform with no invoice attached, and it should be read that way. What makes it worth engaging with anyway is that it matches a shape everyone in this industry has watched before, several times, on a roughly eighteen-month cycle.
What a cold email infrastructure loophole actually was
Here is the honest history. For most of the last four years, cold email at volume has worked because of a gap between what mailbox providers said and what they enforced. The rules were always clear: authenticate properly, keep complaint rates low, do not send unsolicited bulk mail from disposable identities. Enforcement was patchy enough that a well-run operation could sit outside the intent of those rules and inside the tolerance of the filters.
The tooling that grew around that gap is what people mean by infrastructure. Bulk-provisioned workspaces at reseller pricing. Dozens of lookalike domains. Warmup pools that traded artificial engagement between accounts. Rotation that spread volume thin enough to stay under per-domain thresholds. All of it engineering to avoid a limit rather than to earn a reputation.
“Arbitrage is not a moat. It is a countdown with revenue attached.”
That is why the loophole framing rings true. Every one of those techniques works until the provider changes one policy, and then all of them fail at once, because they were correlated the whole time. We have made the same argument in less dramatic terms in shared versus dedicated sending infrastructure: the cheap setup is cheap because someone else is absorbing the risk, right up until they stop.
The number that needs interrogating
Now the 34%. Instantly's 2026 benchmark report, built on their platform data through December 2025, puts the average cold email reply rate at 3.43%. The top quartile clears 5.5%. The top decile clears 10.7%. A 34% positive reply rate is more than three times the elite tier of the largest published dataset in this category.
Cold email reply rates. The first three bars are Instantly's 2026 benchmark report tiers (data through Dec 2025). The fourth is a self-reported, unverified operator claim shown at the same scale for comparison.
That gap has three possible explanations and they are not equally likely. It could be a genuinely exceptional operation, which does happen at small volume with a narrow list and unusual offer fit. It could be a definitional difference, where positive reply is measured against a much smaller denominator such as delivered-and-opened rather than sent. Or it could be a number that grew in the telling.
| EXPLANATION | WHAT WOULD MAKE IT TRUE | HOW YOU WOULD CHECK |
|---|---|---|
| Genuinely exceptional performance | Tiny list, extreme ICP fit, warm-adjacent audience, strong offer | Ask for list size and sent volume alongside the rate |
| Different denominator | Rate measured on delivered or opened rather than sent | Ask which denominator, every time, before quoting anyone |
| Different numerator | Positive counts any human reply including polite declines | Ask what counts as positive and whether out-of-office is excluded |
| Number grew in the telling | No verification available and no incentive to correct | Weight it as marketing, not measurement |
I am not accusing anyone of lying. I am pointing out that 300,000 sends a month at 34% positive replies implies over a hundred thousand interested humans per month, and an agency generating a hundred thousand qualified conversations monthly would be the most famous company in B2B. When a number implies something the world would already know about, the number needs a footnote. We built the reply rate benchmark contradiction around exactly this failure mode, because it repeats constantly.
What the adoption data says is really happening
Underneath the drama there is a slow, boring, well-measured shift, and it is the one that will actually decide who is still sending in 2028. EasyDMARC's 2026 adoption report, published April 10 and covering 1.8 million domains, found 52.1% with a DMARC record, up from 47.7% the prior year. Only 411,935 of those had moved to quarantine or reject. Roughly 9% combined enforcement with reporting.
The split by company size is the interesting part. Fortune 500 domains sit at around 95% adoption with over 80% at active enforcement. Inc. 5000 companies are just over half adopted and largely still at the monitoring-only setting. That gap is the map of where enforcement pressure lands next.
Cold email infrastructure that survives an enforcement wave
There is a reason the rebuild is always described as a rebuild and never as a redesign. When capacity disappears overnight, the instinct is to reconstruct the same machine somewhere the enforcement has not reached yet. New reseller, new domain pool, same architecture. That instinct is what turns a one-time loss into a recurring one, because the thing that broke was not the vendor. It was the assumption that sending identity is disposable.
The setups that came through previous tightenings intact had almost nothing in common with the setups that got wiped out. They were smaller, slower, and boring. They also cost more per email and produced more meetings per email, which is the trade nobody wants to hear about while the cheap version is still working.
That is not an exciting list. It is the list that was still standing after the last three enforcement waves, and it is what we run for clients where a burned domain would be a genuine business problem rather than an inconvenience, including the regulated-sector work behind our FMS investor engagement and across professional services generally.
What to do with your sending setup this month
One more thing worth saying plainly, because the loophole framing can read as fatalism. Cold email is not dying and it is not becoming impossible. It is becoming a channel where the cost of entry is real work rather than clever provisioning, which is what every maturing channel does. Paid search went through it. Display went through it. The operators who complain loudest at each transition are the ones whose advantage was the gap rather than the craft, and the ones who go quiet and keep booking meetings are the ones who had something underneath.
Start by finding out whether you are exposed. Pull the list of every domain you currently send from, check each one for a DMARC record and its policy setting, and mark which ones are still at monitoring-only. If more than a third of your sending volume comes from domains at monitoring-only, you are in the population that the next tightening is aimed at.
Then do the uncomfortable arithmetic. Take your real reply rate, on sent, with out-of-office excluded, and compare it to the 3.43% benchmark rather than to whatever number is circulating this week. If you are below it, the answer is a better list and a better offer, and no amount of infrastructure will substitute for either. If you are above it, protect what you have by moving off anything disposable.
The loophole era ending is not a tragedy for anyone who was not renting one. It is the part of the cycle where the operators who built something dull and durable stop having to explain why their setup costs more. The adoption data behind this is worth reading directly in EasyDMARC's 2026 report.
See where you are cited today
A free snapshot audit of your rankings and AI citations before we ever talk.
Tyler leads work at the intersection of SEO and generative engines at Something Inc., helping B2B brands get ranked and cited across every major AI engine.