Almost every guide on this subject tells you to check your reputation rating in Google Postmaster Tools. Google began retiring that experience in late 2025 — the domain and IP reputation dashboards are gone. The most-cited diagnostic in the field no longer returns the number everyone is telling you to look at.
Providers aren't judging legitimacy
Start here, because it explains the entire puzzle in the title.
Mailbox providers are not assessing whether your email is legitimate, permissioned or well-intentioned. They're predicting one thing: will this recipient want this message? And they infer the answer from observable behaviour, because intent isn't observable.
Which produces the uncomfortable core of the problem. A properly permissioned list of people who signed up two years ago and no longer open anything generates the same signals as a poor-quality list. Low engagement, deletes without reading, occasional complaints. The provider can't see that you did everything correctly. It sees what the recipients do.
The thing that catches good senders Legitimacy isn't a signal. Wanted-ness is — and a list can be entirely legitimate while being demonstrably unwanted.
This is separate from the compliance rules, which are their own gate and covered in our guide to landing in the inbox consistently. You can pass every technical requirement in that piece and still be filtered on reputation, because they're assessing different things.
Reputation isn't one score
The second structural misunderstanding, and it explains why placement varies so sharply between recipient domains.
Reputation operates at several levels simultaneously, and each provider runs its own model weighting them differently.
| Layer | What it covers | Notes |
|---|---|---|
| Domain | Your visible From domain | Now the dominant signal, especially at Gmail |
| IP | The server sending | Weighted more heavily by Outlook |
| Subdomain | Separated sending streams | Lets you isolate marketing from transactional |
| And per-provider: Gmail leans on domain and engagement; Outlook on IP; Yahoo is notably complaint-sensitive. | ||
The practical consequence is one most teams misdiagnose: your reputation can be healthy at one provider and damaged at another at the same time. Gmail delivering fine while Outlook filters isn't a contradiction, it's two different assessments of the same sender.
So before concluding "we have a deliverability problem," segment your placement by recipient domain. If it's uniform across providers, the cause is likely your list or your content. If it's isolated to one, the cause is likely specific to how that provider weights things — and an Outlook-only problem on a shared IP points somewhere quite different from a Gmail-only problem.
Domain reputation travels with you
One consequence worth stating plainly, because it defeats the most common remedy.
Rotating to a new IP rarely fixes a damaged domain. As one practitioner summary puts it, the trust problem travels with the domain — you can move servers, but the name in the From field is what carries the history.
The inverse also matters. A domain with prior history you didn't create — bought secondhand, or previously used for outreach by a team that has since left — arrives with that history attached. Organisations occasionally discover they inherited a reputation problem that predates anyone currently employed there.
Four ways good senders get caught
The specific mechanisms behind "we did nothing wrong."
Recycled spam traps
Spam traps come in two kinds. Pristine traps have never belonged to a real person, so hitting one indicates scraping or purchasing — that's a genuine indictment.
Recycled traps are the danger for legitimate senders. These are real addresses that were abandoned by their owners and later reactivated by anti-abuse organisations specifically to catch senders who never clean their lists.
Which means a perfectly permissioned list becomes hazardous purely through age. Someone who opted in genuinely in 2021, changed jobs in 2023, and whose address was recycled in 2025 is now a trap sitting in a list you acquired entirely properly. No amount of good intent protects you; only list hygiene does.
Campaign Monitor
Email marketing platform with a drag-and-drop builder, segmentation, automation and analytics for campaigns.
Best for: Small & mid-sized teams sending polished email campaigns
Shared IP contamination
On a shared pool, your placement is partly a function of every other sender on it. If a neighbour hits traps or generates complaints, that can affect you even when your own practices are impeccable.
This is worth checking rather than assuming. Ask your platform what your IP situation is, and if you're seeing Outlook-specific problems in particular — given how heavily it weights IP — that's a strong candidate cause.
Volume irregularity
Sending 500 a week for months and then 50,000 in a day looks like a compromised account or a purchased list, regardless of why you actually did it.
This catches seasonal senders hardest. A retailer who sends lightly all year and then floods in November is producing exactly the pattern filters are tuned to catch — which is why volume ramping belongs in peak season preparation in October rather than being discovered in week one of the campaign.
Bounce rates from ordinary list decay
Reported thresholds put under 2% in safe territory. Above roughly 5%, analysis has found inbox placement dropping from around 89% to 62% within thirty days.
Addresses decay naturally — people leave jobs, abandon accounts, change providers. A list nobody has validated in two years will bounce at rates that look like list-buying to a filter, without a single purchased address in it.
Monitoring after the dashboards went away
Google's retirement of the classic Postmaster Tools experience matters practically, not just as trivia. The reputation graphs that showed High, Medium, Low or Bad are no longer there; the current experience is compliance-focused, reporting spam rates, authentication pass rates and delivery errors.
So the field has shifted from reading a rating to computing leading indicators yourself. What to watch instead:
- Spam complaint rate, which the tooling does still report. Keep it under 0.1%; sustained above 0.3% is where blocking starts.
- Bounce rate per send, not averaged over months. A single bad send shows up here immediately.
- Seven-day engagement trend rather than a single campaign's numbers.
- Placement by recipient domain, as above. This is now your best proxy for the ratings that disappeared.
- Microsoft SNDS for how Outlook-side systems view your IPs, including complaint rates and trap hits.
- DMARC report pass rates, which surface authentication drift before it becomes a delivery problem.
Two cautions on reading the data. Aggregate sender scores from third-party tools are lagging indicators — by the time one drops, the damage is already done and you've lost weeks of warning. And a reported "100% spam rate" on a zero-send day is an artefact of delayed complaints against no volume, not a catastrophe; the thirty-day trend is what matters, not any single reading.
It's also worth acknowledging that the underlying measurement got harder some time ago — open-based engagement stopped being reliable when image pre-loading became widespread, as covered in what the privacy updates changed. Between that and the retired dashboards, senders are now diagnosing with visibly worse instruments than they had three years ago.
Damage is asymmetric
The single most useful thing to internalise about reputation, and the strongest argument for prevention.
Reputation falls in days and rebuilds over weeks. Industry guidance puts IP recovery at roughly two to four weeks of disciplined sending, and domain recovery at six to twelve weeks. One bad campaign to an old list can cost you a quarter.
That asymmetry has a planning consequence people rarely draw. If your commercial year has a period that genuinely matters — a launch, a peak season, a funding announcement — the reputation work happens two to three months before it, not during. A sender who damages reputation in September has not recovered by late November.
It also reframes list cleaning. Suppressing unengaged contacts feels like destroying an asset. Against a six-to-twelve-week recovery window, it's cheap insurance against a cost most teams have never priced.
Content signals that trip clean senders
Less influential than reputation but worth eliminating, because these are free fixes that occasionally tip a borderline sender the wrong way.
- Link shorteners. They obscure the destination, which is exactly what abusive senders use them for. Link to your own domain.
- Link domain mismatch. Sending from one domain while every link points somewhere unrelated reads as a forwarding operation.
- Image-only emails. A message that is one large image with no meaningful text can't be assessed on content, and filters treat unassessable mail cautiously.
- Sudden formatting changes. A sender whose mail has looked one way for two years and abruptly looks entirely different is a weak compromise signal.
None of these will sink a strong sender. All of them add friction to a marginal one, and the fixes cost nothing — which makes them worth clearing before you start investigating anything harder.
Subdomain segmentation
The most useful structural protection available, and underused.
Send different streams from different subdomains — marketing from one, transactional from another. This lets each build its own reputation, so a marketing campaign that goes badly doesn't take your password resets and order confirmations down with it.
That separation matters more than it used to, because the compliance thresholds now count all mail from a domain together. A single reputation covering both streams means your most business-critical email inherits the risk profile of your least critical.
Set this up before you need it. Subdomains need their own warm-up period, so creating one mid-crisis adds delay to a situation that's already time-constrained — and any automated flows pointed at the new subdomain need their authentication re-verified rather than assumed.
The list you should be building instead
A structural point that sits underneath all of this.
Every mechanism in this article rewards the same thing: a smaller list of people who genuinely want the mail. Trap risk falls because the list is current. Complaint rate falls because recipients recognise you. Engagement signals improve because the denominator contains fewer people who were never going to open anything.
Which means the acquisition decision determines the deliverability outcome months later. How an address was collected — confirmed opt-in versus a pre-ticked box versus an imported spreadsheet — predicts its behaviour more reliably than anything you do at send time, and that is the argument in building an owned audience properly. Reputation problems are usually acquisition problems that took a year to surface.
Why the environment keeps tightening
Brief, because it explains problems appearing without any change on your side.
Automated outreach multiplied sending volume across the industry, and filters were tuned in response against the patterns automation produces. Paired testing has found AI-written cold email flagged as spam at roughly double the rate of human-written equivalents — the dynamic covered in what autonomous prospecting means for marketers.
The consequence for a legitimate sender is that the bar rose without you doing anything. Teams reporting new deliverability problems in 2026 having made no changes to their own programme are frequently correct — the environment moved, not them.
There's a newer layer worth watching too. As inbox assistants triage, summarise and file mail on recipients' behalf, some of the engagement signals providers rely on are increasingly generated by software rather than people. What that does to reputation modelling isn't settled, but a message read by an assistant and filed without a human ever seeing it produces a different signal from one a person opened and replied to.
What to do
- Split your placement by recipient domain before diagnosing anything. It tells you which of four different problems you have.
- Validate the list before large sends, not after placement drops. Bounces are the fastest route to damage and the easiest to prevent.
- Suppress anyone with no engagement in six months. This removes trap risk and improves the signal for everyone remaining.
- Separate marketing and transactional onto different subdomains, and do it now rather than during an incident.
- Ramp volume gradually before any seasonal peak. Start six to eight weeks out.
- Check whether you're on a shared IP, particularly if Outlook is the outlier.
- Watch complaint rate and bounce rate per send as your primary instruments now that the reputation ratings are gone.
If you've inherited a domain with history you can't account for, or you're seeing persistent blocks after weeks of clean sending, that's the narrow case where moving to a fresh sending domain is genuinely the right call rather than an avoidance of the underlying fix — and it's the point at which specialist email and automation help usually costs less than another quarter of degraded placement.
The short version
Providers aren't judging whether your email is legitimate — they're predicting whether recipients want it, and a properly permissioned but disengaged list produces the same signals as a bad one. Reputation isn't a single score either: it operates at domain, IP and subdomain level, and providers weigh those differently, so you can be healthy at Gmail and blocked at Outlook simultaneously. Split placement by recipient domain to find out which problem you have. Watch for recycled spam traps, which turn an old list hazardous through age alone, and for shared IP neighbours you didn't choose. And note that Google's reputation dashboards are gone, so complaint rate, bounce rate per send and placement by domain are now your instruments — used well ahead of anything commercially important, since domain reputation takes six to twelve weeks to rebuild.
Doing everything right and still landing in the spam folder?
We diagnose reputation problems by provider and rebuild sending so your email reliably arrives.
Explore Our Services →Frequently asked questions
Why is legitimate email still landing in spam?
Because mailbox providers are not judging whether your email is legitimate — they are predicting whether the recipient will want it, and they infer that from behaviour rather than intent. A properly permissioned list of people who no longer open anything produces the same signals as a poor-quality list. Add mechanisms that catch senders who did nothing wrong, such as recycled spam traps in older lists, shared IP pools contaminated by other senders, and inherited history on a reused domain, and it becomes possible to be entirely compliant and still filtered.
Is sender reputation a single score?
No. It operates at several levels — sending IP, sending domain, and subdomain — and each mailbox provider maintains its own model that weighs those inputs differently. Reporting indicates Gmail leans heavily on domain reputation and engagement signals, Outlook weights IP reputation more strongly, and Yahoo is particularly sensitive to complaint rates. The practical consequence is that your reputation can be healthy at one provider and damaged at another simultaneously, which is why placement often varies sharply between recipient domains.
Can you still check your sender reputation in Google Postmaster Tools?
Not in the way most guidance still describes. Google began retiring the classic Postmaster Tools experience in late 2025, and the domain and IP reputation dashboards that displayed a High, Medium, Low or Bad rating are no longer part of it. The current experience is compliance-oriented, reporting spam rates, authentication pass rates and delivery errors. That means the single most-referenced reputation diagnostic no longer returns a reputation rating, and monitoring has shifted toward leading indicators you compute yourself.
How long does it take to recover sender reputation?
Recovery is markedly asymmetric — reputation falls in days and rebuilds over weeks. Industry guidance puts IP reputation recovery at roughly two to four weeks of disciplined sending, while domain reputation typically takes six to twelve weeks. That asymmetry is the practical argument for prevention over remediation, and it also explains why rotating to a new IP rarely solves the problem: where the damage sits on the domain, the trust problem travels with the domain regardless of which server sends the mail.
What are spam traps and how do legitimate senders hit them?
Spam traps are addresses maintained by anti-abuse organisations to identify poor list practice. Pristine traps have never belonged to a real person, so hitting one indicates scraping or purchasing. Recycled traps are the dangerous category for legitimate senders: they are real addresses that were abandoned and later reactivated as traps. Any list containing contacts who stopped engaging years ago can contain them, which means a perfectly permissioned list becomes hazardous purely through age if it is never cleaned.
