Three things this week. The DMARC specification completed its years-long evolution: DMARCbis is now a published RFC — three of them, actually — as of May 20. We covered the approaching finish line last week; this is the final crossing. Separately, Intuit reported that Mailchimp revenue fell year-over-year while the rest of their SMB portfolio grew double-digits, and cut 17% of the company in response — with deliverability engineering reportedly among the affected teams. And Validity published an analysis of Gmail open rate drops reaching 30% or more for some senders, with a root-cause breakdown that points at measurement failure rather than delivery failure. One more item: regional ISPs are exiting consumer email faster than most people have noticed, and the long-term implications for the inbox market are worth paying attention to.

We covered DMARCbis last week as "close enough to read now." On May 20 it crossed the line. Three RFCs published simultaneously:

The structural changes are the same ones previewed in the draft: DNS Tree Walk replaces the Public Suffix List for organizational domain determination; the pct tag is deprecated (replaced by the cleaner t=y/n testing mode); psd and np tags are new for public suffix and non-existent subdomain handling.

What changes with publication: this is now the actual standard. Before May 20, pointing someone at DMARCbis required explaining what a draft is and why it mattered. After May 20, RFC 9989 is the spec. That distinction matters for conversations with compliance teams, customers who ask about your authentication posture, and vendors who need a document to point at.

The backward compatibility story is unchanged: existing v=DMARC1 records continue to work. Your pct=100 record doesn't break anything today. But pct is deprecated, and t is the right mechanism for gradual enforcement rollout going forward. If you have pct=<100 in production records, that's eventual cleanup work.

Word to the Wise: The New DMARC Is Here

On May 20, Intuit announced 3,100 layoffs — 17% of the company's workforce — and the earnings context made it explicit: Mailchimp revenue declined year-over-year while the rest of Intuit's SMB products grew double-digits. The word they used publicly was "rightsize." Deliverability engineering was reportedly among the teams that took cuts.

This is worth noting not as a schadenfreude item but as a signal. Mailchimp has been the entry-level email platform for a generation of small businesses. Revenue declining in a period when email marketing overall is growing points to something structural — competitive displacement from newer ESPs, possible pricing friction, possible product stagnation. Losing deliverability engineers in that environment is a compounding problem: deliverability issues erode sender reputation, which accelerates churn to alternatives.

Intuit acquired Mailchimp for $12 billion in 2021. That context doesn't need commentary.

emailexpert: Intuit Cuts 17% of Workforce as Mailchimp Revenue Declines

Validity published an analysis of widespread Gmail open rate drops showing up across their customer base — some senders down 30% or more quarter-over-quarter. Their investigation points to two mechanisms, neither of which is about inbox placement:

Image prefetch frequency dropped. Gmail has historically prefetched email images on delivery, firing tracking pixels even if the user never opens the email (and, separately, even after they do). That prefetch frequency was reduced — so fewer phantom opens are being recorded, which pulls measured open rates down.

Promotions tab sorting changed. In September 2025, Gmail shifted the Promotions tab from chronological to relevance-based ordering. High-volume senders who relied on subscribers checking Promotions in order lost their implicit open floor. The same emails are being delivered; they're just less likely to be at the top when the tab is opened.

The key finding: real engagement may be unchanged. If your Gmail click rates and revenue per send are holding while open rates fall, this is the likely explanation. The measurement degraded; the channel didn't.

Two caveats: Validity is a deliverability vendor, and their analysis reflects their customer base, not a random sample. The mechanisms are real and documented; the 30% figure is from their data, not Google's. If your open rate story is disconnected from everything else you track at Gmail, this post explains why.

Validity: What's Really Behind Gmail's Open Rate Drop — and What to Do About It

A piece in emailexpert documented something that's been happening gradually enough to be easy to miss: regional ISPs are closing consumer email services at an accelerating pace in 2026. EONI, Ptera, Sparklight, TalkTalk, Plusnet, iPrimus — all either closed their email services or migrated users to third parties this year. Comcast is moving Xfinity email users to Yahoo Mail.

The economics are straightforward. Maintaining authentication infrastructure, abuse filtering, and storage for consumer email is expensive, and the retention value for a regional ISP no longer justifies the cost. Most users who ever had an ISP email address as their primary contact have already switched to Gmail, Outlook, or Yahoo years ago. The ones still on ISP email are being migrated there involuntarily.

The platform effect compounds: as ISP email exits, traffic concentrates further into Gmail, Outlook, and Yahoo. Those three providers already dominate consumer inbox traffic. Every ISP exit increases that share. For senders, the stakes of maintaining good standing with those three keep rising. For the industry, it means fewer independent inbox providers setting independent filtering norms.

The monopolization concern the piece raises isn't new, but the pace has accelerated enough that it's worth tracking.

emailexpert: The 2026 Great ISP Email Retreat

California sues Temu over email practices — Pottish v. WhaleCo (Temu's parent), filed May 4 in Los Angeles, alleges spoofed domains, nonsensical sender addresses, and misleading subject lines designed to bypass spam filters. California B&P Code §17529.5 allows $1,000 per violating email in statutory damages. At commercial email volumes, that's an existential exposure. (emailexpert.com)

99.88% of HTML emails fail accessibility checks — The Email Markup Consortium tested 376,348 emails sent between May 2025 and May 2026. All but eight had serious or critical accessibility failures — statistically zero improvement from 2025's 99.89%. Substack, Shopify, and Beehiiv were specifically named. No email client supported all 37 tested features; Apple Mail led with 34. (emailexpert.com)

That's the week. If something is wrong, reply and tell me — I read every response.