Sender reputation: what mailbox providers measure
Sender reputation is the single biggest factor in whether your mail reaches an inbox, and it is also the least visible. There is no score you can look up for most providers, no appeal process, and no number that transfers between them. Gmail's view of you and Microsoft's view of you are built separately from different evidence.
What you can do is understand which signals feed those judgements, because the signals themselves are not mysterious at all.
Two reputations, not one
IP reputation attaches to the address your mail is sent from. It was the dominant signal for years and still matters for volume senders.
Domain reputation attaches to the domain in your From header and, through DKIM, to the domain that signed the message. It has become the more important of the two, for a simple reason: a sender can change IP address in an afternoon, but changing domain means abandoning their brand.
The practical consequence is that moving to a new IP does not reset a damaged reputation. If the domain is the problem, new infrastructure changes nothing.
Domain reputation is also inherited downwards in a useful way. Mail from news.example.com carries some of the reputation of example.com, which is why separating streams onto subdomains — marketing on one, transactional on another — is standard practice. A bad marketing campaign then damages the marketing subdomain rather than the password-reset mail that must never fail.
The signals that actually move reputation
Complaint rate
The proportion of recipients who click "report spam". This is the most damaging signal, because it is a direct human judgement that your mail is unwanted. Google asks bulk senders to keep the spam rate reported in Postmaster Tools below 0.30%, and recommends staying under 0.10%. Those thresholds are small: on a 50,000-message send, 0.30% is 150 complaints.
Lowering complaints is mostly about expectation management — sending what people signed up for, at the frequency implied, with an unsubscribe link they can find.
Bounce rate
A high proportion of invalid recipients signals that you are not maintaining your list, or did not collect it carefully. It also means a lot of your traffic is wasted effort from the receiving server's point of view. Hard bounces are far more damaging than soft ones; the distinction is covered in hard vs soft bounces, and keeping the rate down is what pre-send verification is for.
Spam trap hits
Mail to an address that exists only to catch senders. A single pristine trap hit can cause a blocklisting, and because traps accept mail silently you get no feedback at all until something breaks. See what a spam trap is.
Engagement
Opens, clicks, replies, moving a message out of spam, adding the sender to contacts. Providers do not publish how they weight these, and open tracking has become unreliable since mail clients began pre-fetching images. Replies and manual "not spam" actions remain strong positive signals precisely because they are hard to fake.
The flip side is negative engagement: deletion without opening, and above all sustained silence from a large share of recipients. Mailing a segment that has not opened anything in two years does not get you back in front of them — it tells the provider that your mail is ignored.
Authentication
SPF, DKIM and DMARC do not improve reputation by themselves, but they make reputation attributable. Without them, a provider cannot reliably tell your mail apart from mail that merely claims to be yours. Since 2024 Gmail and Yahoo have required SPF or DKIM from all senders, and SPF, DKIM and DMARC together from senders of 5,000 or more messages a day; Microsoft introduced equivalent requirements for high-volume senders to its consumer domains in May 2025. These are now entry conditions, not optimisations. See the authentication guide.
Infrastructure hygiene
Valid forward and reverse DNS for your sending IPs, TLS on the connection, correctly formed headers per RFC 5322, a working Message-ID, and — for marketing mail — one-click unsubscribe headers as described in RFC 8058. None of these win you anything. Missing them loses you something.
Consistency of volume
Reputation systems model your normal behaviour. A sender that delivers 2,000 messages a day for six months and then sends 200,000 looks, from the outside, exactly like a compromised account. This is the entire reason warm-up exists.
Check an address before you send to it
Mailcheq runs syntax, MX, disposable, role-account and live SMTP checks in the browser, free.
Verify an address →Shared versus dedicated IP addresses
On a shared IP pool your reputation is mixed with that of every other sender in the pool. Your provider polices the pool, which is why they care so much about your bounce and complaint rates. The upside is that the pool carries established volume, so you benefit from its history rather than starting from nothing.
A dedicated IP gives you a reputation entirely your own — which is only an advantage if you have enough consistent volume to maintain it. A dedicated IP sending a few thousand messages a month never accumulates enough signal to establish itself, and periods of silence erode what little it has. As a rough rule, dedicated IPs start to make sense somewhere in the tens of thousands of messages per week, sent on a regular schedule.
Most senders are better off on a reputable shared pool than on an underused dedicated IP.
The two windows you actually get
Most of reputation is opaque, but two providers give you real data about yourself.
Google Postmaster Tools shows domain and IP reputation bands, spam rate, authentication pass rates, encryption and delivery errors for mail you send to Gmail. You verify ownership of the sending domain via a DNS record. It needs a reasonable daily volume to Gmail before it will show data at all, which is a genuine limitation for smaller senders.
Microsoft SNDS (Smart Network Data Services) reports on traffic from IP addresses you control to Microsoft's consumer domains: volume, complaint rates and trap hits. It is IP-based, so on a shared pool you will not have access — your provider does.
Between them these two cover a large share of most European and North American lists. For everything else you are reading indirect evidence: delivery rates by domain, bounce messages, and whether your own seed addresses land in the inbox.
How reputation moves
Reputation falls quickly and recovers slowly. A single bad send — an old list, a spike in volume, a sudden complaint cluster — can move you from a good band to a bad one within a day. Recovering takes weeks of consistent, low-complaint sending, because the system needs enough new evidence to outweigh the old.
There is no reset button. The things that work are unglamorous:
- Stop mailing the segment that generated the complaints, immediately.
- Reduce volume to your most engaged recipients and rebuild from there.
- Verify the list before you resume, so bounces are not compounding the problem.
- Keep the schedule steady; erratic volume looks like a compromised sender.
- Fix authentication before anything else, so your improved behaviour is actually attributed to you.
And resist the instinct to move to a new IP or a new domain. Changing infrastructure to escape a reputation problem is the pattern that mailbox providers watch for most closely, and it is also the most common reason an email service provider suspends an account.
Spotted an error on this page? Email hello@mailcheq.com and we will fix it.