What Is Sender Reputation, and How Can You Improve It?

Sender reputation is the running score that Gmail, Outlook, and every other mailbox provider keeps on you. It isn't a single number, it isn't published, and nobody at Google will tell you what yours is. But it decides whether your cold email lands in the inbox, in Promotions, in spam, or gets rejected at the SMTP handshake before it exists anywhere at all.

Most people discover it the hard way: reply rates were fine for three weeks, then they fell off a cliff, and nothing about the copy changed. That's reputation. This post explains what it's actually made of, which signals move it, how to see it, and what a realistic repair looks like.

What sender reputation actually is

Mailbox providers receive far more mail than they can inspect deeply. So they use history as a shortcut. Every time you send, they record what happened: did the address exist, did the recipient open it, reply, delete it without reading, or hit "report spam"? Those outcomes get aggregated against identifiers you can't easily change, and the aggregate becomes a prediction — when the next message from this sender arrives, where should it go?

There are three layers, and they carry different weight depending on how you send.

• IP reputation — the score attached to the sending IP address. If you send through Gmail or Microsoft 365, you're on their shared outbound pools and IP reputation is largely out of your hands (and largely excellent). It only becomes your problem if you run a dedicated IP on your own SMTP relay.
• Domain reputation — the score attached to the domain in your From address and your DKIM signature. This is the layer that matters most for cold email. It follows you across mailboxes, across IPs, and across platforms. Burn it and a new mailbox on the same domain won't save you.
• Mailbox/sender reputation — provider-side signals tied to the individual sending address, plus per-recipient personal reputation. Gmail weights what *this specific recipient* has done with mail from you in the past above almost everything else.

There's also a fourth, softer layer: content reputation. Reused subject lines, tracking domains, and link domains get fingerprinted. If the link domain in your footer is in twenty other senders' spam-foldered campaigns, that travels with your message even if your own domain is clean.

The signals that actually move the number

Not all inputs are equal. Ranked roughly by how much damage they do per unit:

Complaints are the nuclear option. At Gmail's stated threshold, one complaint per thousand delivered messages is the line you don't want to cross. If you're sending 500 emails a day to a list and two people a day hit report spam, you are at 0.4% and your domain reputation will degrade within a week or two.

Hard bounces are the second-fastest way down. A list with 8% invalid addresses tells the provider you don't know who you're mailing — which is exactly the profile of a bought list. This is why verifying every address before send is non-negotiable rather than a nice-to-have.

Volume spikes matter more than people expect. Going from 40 emails a day to 400 overnight on a domain with three months of history looks like a compromised account, not a growing business. Providers respond to that pattern faster than they respond to content.

How to actually see your reputation

You can't see a score, but you can see proxies. Three tools, in order of usefulness for cold email:

• Google Postmaster Tools — free, requires a DNS TXT record to verify your domain. Shows domain reputation (Bad / Low / Medium / High), IP reputation, spam rate, and authentication pass rates. Caveat: it only populates once you're sending meaningful daily volume to Gmail addresses — typically a few hundred a day. Most cold email senders never hit the threshold on a single domain, which is an argument for consolidating rather than spreading across ten throwaway domains.
• Microsoft SNDS / JMRP — IP-level data and a complaint feedback loop. Only useful if you control your sending IP.
• Seed inbox-placement tests — you send a campaign to a set of addresses you own across Gmail, Outlook, Yahoo, and corporate Microsoft 365 tenants, then check where each one landed. This is the only method that tells you about *placement* rather than about upstream signals. It's a sample, not truth, but a run where 4 of 10 Gmail seeds land in spam is a real signal.

What you shouldn't rely on: open rates. Security scanners and image proxies fire pixels on messages nobody read, so a "healthy" 60% open rate can sit on top of a domain that's already spam-foldered. Replies and verified link clicks are the only engagement numbers worth trending.

How to improve a reputation that's already damaged

Reputation is asymmetric. It takes weeks of good behavior to build and about four days of bad behavior to wreck. Repair is slow and boring, and the order matters.

1. Stop sending from the affected domain. Not slow down — stop, for at least 72 hours. Every additional message sent into a bad reputation state reinforces the classification. This is the step people skip because pausing feels like losing pipeline.

2. Fix authentication first. SPF, DKIM, and DMARC all passing, on every mailbox, with alignment. If DKIM is failing on 5% of sends because a relay is rewriting headers, nothing else you do will hold. Our plain-English setup guide covers the exact records.

3. Clean the list harder than feels necessary. Re-verify everything. Drop catch-all domains you can't confirm. Suppress anyone who's been mailed three times without a reply — repeated non-engagement is itself a negative signal.

4. Cut copy risk. Remove image-heavy templates, shorten links, drop the tracking pixel entirely if you don't need it, and run drafts through a spam scorer before send. A SpamAssassin score above 3 is worth rewriting.

5. Restart at 10-15% of previous volume and ramp over 3-4 weeks. Not a warm-up tool substitute — actual campaign volume, ramped. Send the best-fit 200 leads you have, the ones most likely to reply. Positive replies are the fastest reputation currency there is.

6. Only then consider a new domain. New domains are not a fix, they're a reset with a 4-8 week cost. If you move to a fresh domain without changing the list quality and copy that burned the first one, you'll burn the second one on the same schedule.

Realistic timeline: a Low reputation domain with clean fundamentals and a disciplined ramp usually returns to Medium/High in three to five weeks. A Bad reputation domain with trap hits may never fully recover, which is the honest case for retiring it.

The habits that prevent this

Preventing reputation damage is mostly about caps and stop rules, not tactics.

• Cap daily volume per mailbox at 30-50 for cold sends, regardless of what the mailbox "can" handle.
• Keep a per-campaign bounce kill switch — if bounces exceed 3% in the first 100 sends, pause automatically.
• Treat unsubscribes as a gift. One-click unsubscribe on every message converts would-be complaints into harmless opt-outs.
• Maintain a global suppression list across every campaign and mailbox, not per-campaign.
• Watch the trend line, not the daily number. Three consecutive days of declining reply rate on stable copy is a reputation warning, not a copy problem.

Where Sendvanta fits

Most of the work above is monitoring you'd otherwise do by hand across four dashboards. Sendvanta scores reputation risk in layers — mailbox, domain, DKIM, and IP — and automatically slows or pauses sending when a threshold is crossed, so a bad list doesn't get to run for three more days before you notice. It also runs SpamAssassin scoring on drafts before send, supports seed-mailbox inbox placement tests, classifies engagement bot-aware so you're trending real replies rather than scanner opens, and applies suppression and one-click unsubscribe by default. Mail goes through your own Gmail, Microsoft 365, or SMTP mailboxes — never a shared pool, so the reputation you build is yours. You can see the full feature set at /features, and the free plan is $0 forever with 1,000 active leads and 3,000 emails a month if you want to test the monitoring on a live campaign before paying anything.