What’s the difference between infrastructure and brand-level reputation?

Infrastructure reputation belongs to the boxes and IPs doing the actual SMTP work. Brand reputation belongs to your domain. Mailbox providers track both, but they weight them differently, and the trend over the last decade has been clear: the domain matters more than the wires.

Infrastructure reputation is everything tied to the sending plumbing. The IPv4 or IPv6 address that opens the connection. The PTR record on that IP. The HELO/EHLO hostname your MTA announces. The ESP or relay you rent. If you send through SendGrid today and switch to Mailgun tomorrow, your IP changes, your PTR changes, the shared pool changes. Infrastructure reputation can be swapped out in an afternoon by changing DNS and your sending account.

Brand reputation belongs to the domain in the From: header, and to a lesser extent the DKIM signing domain (d=) and the SPF Return-Path domain. Mailbox providers attach a long-term trust score to that identity. Move from SendGrid to Mailgun and acme.com is still acme.com. The complaint rate, the spam trap hits, the engagement curve from your last two years all follow the domain.

This is why Gmail and Yahoo's 2024 bulk sender rules forced a DMARC-aligned From domain. They want a stable identity to score. An IP can be rotated by your ESP every Tuesday. A domain is the thing you put on business cards. Read the actual requirements at Google's sender guidelines and you'll see almost every threshold (the 0.30% spam complaint cap, the one-click unsubscribe, the DMARC policy) is keyed to the domain, not the IP.

What this looks like in practice

A few concrete scenarios from list cleaning work:

• Client switches ESPs hoping to escape a block. They were on Mailchimp, getting filtered into Gmail Promotions. They move to Klaviyo. Same domain, same list, same content. Result: same placement. The IP changed. The reputation didn't.
• Client buys a new domain to "start fresh." Now they're sending from acme-mail.com instead of acme.com. Gmail treats it as a cold domain with no history. Inbox rate climbs for a few weeks, then craters once the same bad list habits show up in the engagement data. Brand reputation is harder to outrun than people think, because the behavior is what builds it.
• Shared IP pool gets one bad neighbor. Your IP reputation tanks for a week. Your domain reputation barely moves. You ride it out and recover. This is the textbook case for why domain reputation replaced IP reputation as the primary signal.

Which one breaks deliverability faster

Infrastructure problems are loud and fast. A blocklisted IP, a broken reverse DNS, a misconfigured SPF: you see bounces within hours. Brand problems are slow and quiet. Spam folder placement drift over six weeks. Lower open rates on the same subject lines you used last quarter. A creeping complaint rate. These don't bounce. They just filter.

That's why the two get monitored differently. For IPs, check Talos and SenderScore. For the domain, check Gmail Postmaster Tools and Microsoft SNDS. For a full breakdown of where to look, see tools that measure domain reputation.

The practical rule

Don't try to escape a reputation problem by changing infrastructure. The list, the content, and the sending patterns made the reputation. Fix those first. Switching ESPs without fixing the underlying issue just buys you a few weeks of grace before the same domain hits the same filters at the new provider.

If the domain is genuinely poisoned, the answer isn't a new IP. It's segmenting traffic onto a separate subdomain with its own warmup, which is a different topic covered in how to isolate high-risk sends by subdomain. Infrastructure is the vehicle. The domain is who you are.