Sender reputation is memory, not a score.
Mailbox providers do not meet each campaign for the first time. They remember what happened when messages with the same identities arrived before.
By Moosewave
Published · 8 min read
The short version
Reputation is the past leaning on the next send.
- 01There is no universal sender-reputation number shared by every mailbox provider.
- 02Domains, IP addresses, subdomains, and message streams can carry overlapping histories.
- 03Complaints, unwanted volume, invalid recipients, and inconsistent patterns can make later mail harder to place.
- 04Recovery comes from changing the behavior that creates the evidence, then accumulating better evidence over time.
The imaginary score
People like scores because scores collapse a system into a decision. Seventy-two is better than forty-nine. A green badge is safer than a red one. The interface appears to have finished the thinking for us.
Sender reputation resists this treatment. Gmail exposes domain and IP reputation through its own Postmaster Tools. Yahoo and Microsoft make their own filtering decisions. A third-party monitoring product can summarize signals it can observe, but it cannot reveal a single master score because there is no single master judge.
Even inside one provider, reputation is not necessarily one property attached forever to a company name. Different sending domains, subdomains, IPs, streams, recipients, and recent patterns can create different contexts.
A score says what you are. Memory records what you keep doing.
Yesterday’s field note explained that authentication proves identity, not placement. Reputation is what makes that identity operationally useful. Once a provider can associate mail with a stable sender, it can learn from the history of that stream.
What carries the memory
Asking whether a sender has “good reputation” is incomplete until you ask: attached to what, and judged by whom?
| Identity | What it groups | Important boundary |
|---|---|---|
| Domain | Mail associated with an authenticated or visible domain | A new IP does not erase domain-level history |
| Subdomain or stream | A separable class such as transactional or marketing mail | Separation helps only when the behavior is genuinely separate |
| IP address | Mail sent through a particular infrastructure endpoint | Shared-IP history can include other senders |
| Recipient relationship | How particular recipients respond to a sender or stream | Aggregate health does not predict every mailbox |
This is why infrastructure advice needs context. A dedicated IP gives one sender control over that IP’s behavior, but a quiet dedicated IP also lacks the stable volume needed to build a useful history. A well-managed shared pool can be healthier than an underused dedicated address. Ownership is not the same as reputation.
The better architecture separates mail by meaning. Password resets, receipts, product notices, and promotions create different recipient expectations. Streams should be distinct enough to diagnose without becoming a maze of identities no one can operate.
What mailbox providers remember
Providers do not publish their complete filtering models, and responsible advice should not pretend otherwise. Their public guidance is still consistent about the kinds of behavior senders control.
Google says frequent user spam reports can lower domain reputation and make future messages more likely to be marked as spam. It recommends sending only to people who subscribed, keeping complaint rates low, increasing volume gradually, avoiding bursts, and monitoring reputation and server responses. Yahoo warns that sending to people who do not read the mail or who report it as spam harms delivery metrics and reputation.
The list is less mysterious than the word “reputation” makes it sound. Providers observe whether recipients appear to want the mail, whether addresses are valid, whether volume behaves predictably, whether authentication is stable, and whether the sender makes leaving easy.
Opens are not a perfect definition of interest. Privacy protections and automated image fetching distort them. But the operating principle survives: repeatedly sending mail to people who give no useful sign of wanting it is a poor way to teach a provider that the next message belongs in view.
Audience decisions therefore become deliverability decisions. Read our guide to segmentation and suppression alongside the guide to measurement after privacy-protected opens.
Recovery is new evidence
A damaged reputation tempts teams into dramatic resets: buy a new domain, move to a new IP, change providers, or send a huge “re-engagement” campaign. Each can create more uncertainty while preserving the behavior that caused the problem.
A more reliable recovery sequence is quieter.
- 01
Find the affected identity and provider
Segment acceptance and observed placement by mailbox provider, authenticated domain, stream, and sending IP. Do not turn a localized problem into a site-wide migration.
- 02
Stop adding bad evidence
Suppress invalid, unconsented, chronically inactive, and complaining recipients. Pause the source or stream that cannot explain why its audience should expect the message.
- 03
Stabilize authentication and volume
Keep SPF, DKIM, DMARC, DNS, and routing correct. Send at a consistent rate to recipients with the clearest recent evidence of interest; avoid sudden spikes.
- 04
Make leaving easier than complaining
Use visible, functioning unsubscribe controls and honor requests promptly. A person who can exit cleanly does not need to teach the provider with the spam button.
- 05
Watch the trend, not a magic date
Compare provider responses, complaints, bounces, and observed placement over multiple sends. There is no universal number of days after which history resets.
Use the interactive Moosewave walkthrough to see placement evidence in campaign context, and the deliverability guide for the wider system behind these checks.
The product consequence
A product should help people reason about reputation without pretending to possess the provider’s private answer.
That means showing the identities that carried a send, authentication and alignment results, volume changes, audience eligibility, complaints, bounces, provider responses, and observed placement beside one another. The report should make it possible to ask which history changed before it proposes a cure.
It also means refusing a fake universal grade. Moosewave can show evidence and trends. It can identify a provider-specific placement shift. It can connect that shift to changes in audience and sending behavior. It cannot read an unpublished filtering model or guarantee the next destination.
Our security boundaries use the same discipline: state the control, the evidence, and the limit together. In deliverability, that discipline is not modesty. It is what keeps a plausible dashboard from causing an expensive decision.
Reputation feels abstract only when history is hidden. Put the history back beside the result, and the next action gets much easier to see.
Frequently asked questions
Direct answers to the questions a universal reputation score cannot settle.
What is email sender reputation?
Email sender reputation is a mailbox provider’s evolving assessment of mail associated with sending identities such as domains and IP addresses. Authentication, complaints, bounces, sending patterns, recipient response, and other signals can contribute. There is no single reputation shared by every provider.
What is the difference between domain and IP reputation?
Domain reputation is associated with authenticated or visible domains; IP reputation is associated with the sending infrastructure. A campaign can inherit signals from both, and providers may evaluate them differently. On a shared IP, other senders can affect the IP-level history.
How long does sender reputation take to recover?
Mailbox providers do not publish a universal recovery clock. Recovery depends on the cause, the provider, the scale of the damage, and the quality and consistency of later traffic. Treat it as a sequence of better evidence, not a countdown.
Will switching to a new IP fix poor sender reputation?
Not by itself. A new IP has little or no history and must establish one, while domain, audience, and complaint problems can follow the sender. Moving the same behavior to new infrastructure usually postpones the diagnosis instead of solving it.
Sources & method
Provider references
Provider documentation supports the operating claims. The memory model and recovery sequence are Moosewave’s synthesis; they are not a claim to know a provider’s private algorithm.
- Gmail Help — Email sender guidelines. Covers domain and IP reputation, complaints, subscriptions, gradual volume changes, and monitoring.
- Gmail Help — Postmaster Tools dashboards. Describes Gmail’s domain and IP reputation dashboards and their provider-specific scope.
- Yahoo Sender Hub — Sender best practices. Documents authentication, recipient response, complaint control, list hygiene, and unsubscribe expectations.
- Microsoft — Outlook requirements for high-volume senders. Pairs authentication with list hygiene, bounce management, consent, truthful headers, and opt-out paths.
Last reviewed 1 August 2026. Provider requirements change; verify the linked guidance before making an operational change.
Continue reading
Give the memory a stable identity
The previous field note explains what SPF, DKIM, and DMARC prove—and why all three can pass before placement still changes.