Four policy changes, two from email providers and two from phone carriers, turned what used to be deliverability friction into a hard wall.
Between Feb 2024 and Sept 2025, email and phone, the two channels outbound depends on most, went through their most aggressive infrastructure tightening in a decade. Email providers stopped sending non-compliant mail to spam and started rejecting it outright. Carriers moved from user reports to AI scoring in milliseconds. Apple turned every iPhone into a screener. The playbook your team built between 2018 and 2023 was built for an inbox and a phone that no longer exist.
For the first time, the two providers behind most consumer inbox volume aligned on the same enforcement rules. Spam complaint rate became a number you can die from.
Senders pushing more than 5,000 messages per day to Gmail or Yahoo personal addresses must now authenticate with SPF, DKIM, and DMARC, honor one-click unsubscribe within 48 hours, and keep complaint rates below 0.3%. A sender delivering 10,000 emails needs only 30 spam reports to cross the line. For cold outreach, that is a Tuesday.
"This is nothing new. We have always looked at these spam rates. If you're a good sender, your spam rates will be well below 0.3%."
Marcel Becker, Yahoo
The 2024 rules came with a soft glide path. By late 2025, the glide path was gone.
The three largest consumer email providers, Gmail, Yahoo, and Outlook, now share the same baseline. Microsoft originally planned to route non-compliant messages to junk; ten days before launch, it switched to outright rejection. Google followed in November 2025, ending the soft enforcement era. Non-compliant email no longer lands in spam. It does not arrive at all.
The mental model that needs to die: "It probably went to spam." For non-compliant senders at scale, the message simply does not exist.
Hiya powers AT&T. TNS powers Verizon. First Orion powers T-Mobile. Three independent AI engines now make spam decisions before your phone rings, and STIR/SHAKEN does not bypass them.
STIR/SHAKEN proves you are not spoofing your number. It says nothing about whether your behavior looks like spam. Carriers evaluate volume per number, call duration, answer rates, and complaints. An SDR making 80 calls per hour with low connect rates and short calls looks identical to a robocaller. Industry guidance now recommends staying under 75 calls per day per number, a number that would have been laughable in 2022.
15% of businesses surveyed reported losing more than $100,000 in revenue from incorrect spam labeling. The number that picks up a "Spam Likely" call is functionally zero.
iOS 26 launched in September 2025 with system-level call screening. 150 million U.S. iPhones are now potential gatekeepers, and branded calling does not bypass it.
An unknown number is intercepted before the phone rings. An automated voice asks the caller to state their name and reason; the response is transcribed live; the recipient decides whether to answer. Branded Calling ID displays during the screening interaction but does not bypass it for callers without prior contact, which means every cold caller, by definition. Repetitive dialing makes things worse: iOS 26 watches for aggressive call patterns and silences future calls automatically.
For a decade, outbound teams operated on a tacit assumption: the channel is open, and getting through is a function of effort. More dials, more sequences, more touches. That assumption no longer holds. The pipe has gatekeepers now: four of them, two on email and two on phone, all AI, all scoring behavior in milliseconds, none of them caring about your intent.
Teams whose numbers are falling are not failing at outbound. They are succeeding at an outbound motion the infrastructure no longer permits. The next three whitepapers in this series cover what that means in practice: what is dead, what is working, and what the operating model looks like when relevance and reputation, not volume, are the only signals that get a message through.