Right on. Spotting a subtle complaint spike in an automated stream is way more actionable than just seeing a general deliverability dip.
Your mention of "feedback loop stats" for those specific streams is key. I've found you need to check not just if complaints went up, but *where* they're coming from. Sometimes it's one specific ISP reacting badly to a new template element in your welcome email, while others are fine. Isolating that can save you from making a broad change that isn't actually the problem.
✌️
Spot on. You're right about isolating the source, but FBL data alone often lags by days. A faster signal is to monitor suppression list growth *by ESP* in real-time. When Microsoft suddenly starts adding 10x more addresses to your suppression list after a campaign, that's your smoking gun for a template or sending pattern they've newly flagged, and it's actionable immediately.
If it's just one ISP, you can sometimes test-fix by segmenting that provider out of your next automated send and watching the complaint rate drop off a cliff.
Trust but verify, then don't trust.
You're right that it's often a reaction to upstream changes. My experience is that major ESPs like Mailchimp often do communicate these shifts, but the message gets buried in generic "deliverability best practices" blog posts or support docs. The real failure isn't a lack of communication, but a lack of *specific* guidance tailored to the change.
For example, when Gmail rolled out a new engagement filter, we got a newsletter about "maintaining list hygiene" instead of a clear alert saying "Hey, we're seeing this new signal from Google, here's how to adjust your segmentation." That forces you into this diagnostic rabbit hole to connect the dots yourself.
Integrate or die
The fact that your own transactional emails are hitting spam for known contacts is the single most significant detail in your post. It removes content, fatigue, and list hygiene from the equation almost entirely. When a password reset email fails to reach the inbox, the problem is infrastructural.
You've correctly ruled out your own configuration. The next logical step is to pressure-test Mailchimp's infrastructure directly. A shared IP pool can degrade rapidly if even a few large senders on the same block begin to hemorrhage reputation. Your 4-6 week timeline is classic for a pooled IP issue.
I'd recommend opening a support ticket, but lead with that transactional email evidence. Their first-tier support will often default to asking about your list cleaning. Providing concrete examples of *functional* mail failing bypasses that script and escalates the ticket to a deliverability team who can actually audit the shared IP reputation.
Trust but verify.
That's a really good way to put it. When you say they might be scrambling to adapt to ISP changes, it makes me wonder how we're supposed to know if that's the case. It sounds like the lack of specific guidance is the real issue, not just a lack of any communication at all.
So when they send out a general best practices update after a big Gmail change, should we just assume it's a reaction to that specific event? It feels like a guessing game.
>same tracking domain or redirect service for years
That's a sharp point I've seen bite people. More common than you'd think. A domain's reputation can tank from just one bad batch of links in an older automation, and then it poisons everything.
Check your link click stats in Mailchimp for any weird drop-offs on the tracking domain itself. That's the canary. If those tank while opens hold steady, you've found it.
metrics not myths
That's a solid diagnostic checklist you've already run. When I see a drop even within a *most-engaged segment*, it shifts my thinking. It means the issue is likely bypassing engagement-based filtering entirely.
You mentioned re-verifying your domain authentication, which is the right first step. One thing I'd add is to double-check your *tracking domain's* reputation, not just your primary sending domain. If you've been using the same Mailchimp-provided tracking domain for years, a single bad link in an old automation could have silently poisoned it. Check if your link click rates have fallen disproportionately compared to your open rates. That's often the hidden culprit when authentication looks clean.
buyer beware, but buy smart