Mechanics of Domain Reputation at Gmail

• September 2021
• Engineering Memo · External Release

Gmail inbox placement for high-volume senders is determined primarily by domain reputation — a per-domain signal that Gmail's systems maintain based on user behaviour with messages signed by that domain's DKIM key. Unlike IP reputation, which is visible in Postmaster Tools on a per-IP basis, domain reputation aggregates signals from all IPs and sending sources that use the same DKIM signing domain. Understanding how Gmail builds this reputation, what signals move it, and how to measure its current state is the foundational knowledge for managing Gmail deliverability at scale.

How Gmail Builds Domain Reputation

Gmail's domain reputation system works by attributing user behaviour to the DKIM signing domain in each message. When a Gmail user receives a message signed by example.com's DKIM key and marks it as spam, that action is recorded as a negative signal against example.com's domain reputation. When a user opens, clicks, or replies to a message from example.com, those actions are recorded as positive signals. Over rolling time windows — typically 7, 14, and 30 days — Gmail aggregates these signals into a domain reputation tier: High, Medium, Low, or Bad.

The reputation tier is not a score — it's a classification that maps to specific delivery treatment. High reputation domains receive the most generous inbox placement and rate limit treatment. Medium reputation domains see more cautious treatment, with higher proportions of messages being routed to spam or subject to more aggressive content filtering. Low reputation domains face aggressive spam filtering. Bad reputation domains are typically blocked entirely, with messages rejected at the SMTP layer rather than accepted and filtered.

The weighting of signals is not published by Google, but operational observation across high-volume environments consistently shows that negative signals (spam markings and hard bounces) have substantially more weight per event than positive signals (opens and clicks). A complaint rate of 0.1% — one spam marking per 1,000 messages — is sufficient to produce measurable reputation degradation within weeks. It takes substantially more positive engagement to offset each complaint received.

Reading Domain Reputation in Postmaster Tools

Gmail Postmaster Tools provides direct visibility into domain reputation through its Domain Dashboard. After registering and verifying the sending domain (by adding a TXT record to DNS), the dashboard shows the daily domain reputation tier and the daily spam rate — the percentage of Gmail users who marked messages from the domain as spam. Both signals update daily and reflect the previous 24 hours of sending activity (with a typical 24–48 hour reporting lag).

The spam rate graph is the most operationally actionable signal in Postmaster Tools. The tier classifications (High/Medium/Low/Bad) update less frequently and represent a smoothed view of the domain's reputation state. The spam rate is the leading indicator that precedes tier changes: a spam rate trending above 0.1% will typically produce a tier degradation within 2–4 weeks if it persists. Monitoring the spam rate trend — not just the current value — is what enables early intervention before tier changes occur.

The IP reputation tab in Postmaster Tools shows per-IP reputation for IPv4 addresses that have sent sufficient volume to Gmail in the measurement period. This per-IP view complements domain reputation by showing whether specific IPs in the pool are dragging the domain reputation downward. An IP showing Low reputation in Postmaster Tools while the domain is at Medium suggests that the Low-reputation IP's signals are contributing to the domain's failure to reach High reputation — addressing the per-IP issues often has direct domain reputation benefit.

Domain Reputation vs IP Reputation: Which Matters More

For high-volume senders with established sending history, domain reputation is the dominant inbox placement signal. A domain with High reputation sending from a new IP will experience much better inbox placement than a domain with Medium reputation sending from a High-reputation IP. The domain reputation provides the signal foundation; IP reputation is a secondary adjustment within the range the domain reputation establishes.

This hierarchy has important practical implications. Adding new IPs to a pool does not significantly improve inbox placement if the domain reputation is the constraint — the new IPs will be subject to the same reputation ceiling as the existing IPs. Improving domain reputation (by reducing complaint rates, improving engagement rates, and maintaining consistent clean sending) produces inbox placement improvements that no IP investment can replicate. IP additions produce throughput capacity improvements; domain reputation improvements produce inbox placement improvements. Both matter, but they address different variables.

Conversely, a single IP with poor reputation in a large pool is unlikely to significantly degrade overall domain reputation if the pool's aggregate signals are positive. The domain reputation is built on the weighted aggregate of all sending from that domain, not on the worst-performing IP in the pool. However, an IP with very poor reputation (Bad in Postmaster Tools) can add sufficient negative signals to drag the domain reputation downward — which is the operational case for monitoring per-IP reputation and retiring or re-warming IPs whose signals are clearly negative.

How Domain Reputation Recovers

Domain reputation recovery follows the same rolling-window logic as reputation building. When the underlying causes of reputation decline are addressed — complaint rates reduced, bounce rates cleaned up, engagement-based suppression implemented — the positive signals from subsequent clean sending begin replacing the negative signals from the period of poor-quality sending in the rolling window. Recovery is typically visible as a spam rate decline in the Postmaster Tools spam rate graph within 2–3 weeks of consistent clean sending, followed by a tier improvement within 4–8 weeks.

The recovery timeline can be compressed by increasing the proportion of highly engaged contacts in the sending mix during the recovery period. Sending exclusively to the 30-day engaged segment — contacts who opened or clicked in the past 30 days — during the recovery period maximises the positive signal density per campaign. Each send to this segment generates a high proportion of opens and a very low proportion of complaints, accelerating the signal replacement in Gmail's rolling window.

What does not accelerate recovery: switching IPs, changing sending domains, or adding new IPs. Since domain reputation is attributed to the DKIM signing domain rather than the IP, these changes do not affect the reputation signal that needs to improve. Only the quality of sending from the affected domain drives recovery. Operators who try to "escape" a damaged domain reputation by moving to a new domain will find that the new domain must build its reputation from scratch, which takes time — and may be watched more carefully by Gmail's systems if the sending volume and patterns resemble the previous domain's history.

Maintaining High Reputation: The Operational Practices

High domain reputation at Gmail is maintained through consistent operational practices that keep the underlying signal inputs — complaint rate, bounce rate, engagement rate — within the ranges that High reputation requires. The specific practices: complaint rate below 0.05% (target; Gmail's published threshold for concern is 0.10%, but operating below 0.05% provides a buffer against occasional spikes), hard bounce rate below 0.3% (maintained through acquisition-time validation and real-time bounce processing), and engagement rate sufficient to generate positive signals on each campaign (maintained through engagement-based list segmentation and inactive contact suppression).

Authentication is a prerequisite for domain reputation to be attributed correctly. Messages that fail DKIM signature verification, or that have misaligned DMARC domains, may not be attributed to the intended domain in Gmail's reputation model. SPF, DKIM, and DMARC must all be correctly configured and passing before domain reputation management is meaningful — without correct authentication, there is no reliable attribution of sending signals to the intended domain.

Volume consistency also contributes to reputation stability. Programmes that send consistent weekly volumes establish a reliable baseline that Gmail's systems can evaluate against. Programmes with highly variable volume — sending nothing for weeks, then large campaigns — give Gmail's systems less data for reputation assessment and may face more conservative treatment during irregular sends than consistently-sending programmes at equivalent complaint and bounce rates. Building a consistent sending cadence is not just an engagement best practice — it contributes directly to the stability of the domain reputation signal that determines inbox placement.

Domain reputation is the most important single deliverability variable for Gmail, and Gmail is typically the largest ISP destination for most consumer email programmes. Managing it correctly — understanding its mechanics, measuring it through Postmaster Tools, and maintaining the operational practices that keep it at High — is the foundational work that makes everything else in email deliverability management effective. Without High domain reputation at Gmail, no infrastructure investment, content optimisation, or technical improvement can fully compensate for the inbox placement ceiling that lower tiers impose.

The DKIM Signing Domain as Reputation Anchor

Gmail attributes domain reputation to the d= value in the DKIM-Signature header — the domain specified in the DKIM signing configuration, which may or may not be the same as the From: header domain. For DMARC alignment, the DKIM signing domain must match the From: domain (or a subdomain of it). When this alignment is correct, user behaviour (spam markings, opens, clicks) is attributed to the From: domain, building domain reputation where it is most visible to senders monitoring Postmaster Tools.

A common misconfiguration: a programme that uses a third-party email service provider has its DKIM signed by the ESP's domain rather than its own domain. The reputation built through this sending is attributed to the ESP's domain, not to the sending organisation's domain. When the organisation migrates to its own infrastructure with its own DKIM signing, it starts with no domain reputation at Gmail — the reputation built through the ESP is not portable. This is one of the most significant, and least anticipated, consequences of ESP-to-own-infrastructure migrations: the domain reputation must be rebuilt from scratch.

Subdomains have independent domain reputation from the root domain. A programme that sends marketing email from marketing.example.com while transactional email goes from mail.example.com maintains separate reputation signals for each subdomain. This isolation is by design — the reputation for one subdomain does not directly contaminate the other. However, Gmail's systems are aware of the relationship between subdomains and root domains; a root domain with Bad reputation may affect how Gmail treats messages from its subdomains, and very poor subdomain reputation can affect how Gmail evaluates the root domain.

Gmail's 2024 Bulk Sender Requirements: Domain Reputation Context

Gmail's bulk sender requirements, enforced from February 2024, formalised several practices that were previously best practice recommendations into explicit requirements for senders above specific volume thresholds. The requirements most directly relevant to domain reputation management: one-click unsubscribe via RFC 8058 List-Unsubscribe-Post, spam rate maintenance below the published thresholds (0.10% as a soft threshold, 0.30% as a hard threshold above which delivery is affected), and DMARC with at least p=none for the sending domain.

These requirements did not change how Gmail's domain reputation system works — they formalised the signal thresholds and authentication requirements that the reputation system had already been using. Programmes that were already maintaining spam rates below 0.10% and operating with correct authentication were not significantly affected. Programmes that were operating above the published thresholds or without authentication saw enforcement consequences that had previously been applied inconsistently.

The one-click unsubscribe requirement is particularly relevant to domain reputation because it reduces the complaint rate. When recipients who want to stop receiving email can unsubscribe with a single click, the proportion who choose to mark as spam instead (because unsubscribing is difficult) decreases. Lower complaint rates from easier unsubscribe mechanics directly benefit domain reputation. The requirement is an alignment of technical mandate with reputation self-interest: implementing one-click unsubscribe correctly is both required and beneficial for domain reputation health.

Reading Domain Reputation Trends for Operational Decisions

The operational value of Postmaster Tools domain reputation data is in trend analysis, not snapshot reading. A single day's spam rate of 0.07% is informative but not actionable on its own — is it a spike from one campaign or a sustained trend? The Postmaster Tools graph view over 30 or 90 days shows the trend that makes the single-day value meaningful: a rate that has been consistently below 0.05% for 6 weeks and spiked to 0.07% yesterday warrants investigation but not alarm. A rate that has been climbing from 0.04% to 0.07% over three consecutive weeks warrants both investigation and intervention before it reaches the 0.10% threshold where reputation degradation becomes likely.

The campaign correlation is the next analytical step after trend identification. When the spam rate spikes in the Postmaster Tools data for a specific day, the accounting log for that day shows which campaigns were delivered to Gmail recipients. Correlating the spike timing with specific campaign sends identifies which campaign generated the elevated complaint signal. The investigation then focuses on the list segment used for that campaign, the content characteristics that may have elevated complaint rate, and the sending frequency relative to previous sends to that segment.

This trend-and-correlation workflow is the daily operational practice that keeps domain reputation issues from accumulating undetected. It takes 10–15 minutes per day: check Postmaster Tools spam rate, compare to previous days, and if elevated, correlate with yesterday's campaigns in the accounting log. The investment is modest; the early detection it provides is the difference between catching a reputation problem at the signal stage and discovering it at the inbox placement stage — weeks later, after the reputation has already degraded.

Domain reputation at Gmail is not a mystery to manage or a black box to observe passively. It is a signal system with documented inputs (complaint rate, bounce rate, engagement signals, authentication), published measurement tools (Postmaster Tools), and known operational levers (list quality, sending consistency, content relevance, unsubscribe accessibility). Understanding the mechanics and applying them systematically produces domain reputation that reflects the programme's actual quality — and the inbox placement outcomes that High domain reputation delivers over the lifetime of the programme.

Diagnosing Domain Reputation Problems: A Systematic Approach

When domain reputation declines — visible as a tier change in Postmaster Tools or a sustained spam rate increase — the diagnosis follows a structured path. First, identify the timeline: when did the spam rate begin rising? Postmaster Tools shows daily data going back 90 days; the onset date determines which campaigns and list changes occurred immediately before the rise began. Second, isolate the cause: was the spam rate increase accompanied by a bounce rate increase (list quality problem), or was the spam rate isolated (content or frequency problem, or specific list segment problem)? Third, identify the specific campaign or list segment: which of the campaigns sent in the days before the onset generated the most complaints, based on accounting log correlation?

The diagnosis almost always points to one of four causes: a specific list segment with elevated complaint propensity (recently acquired contacts from a low-quality source, or contacts re-engaged without proper warming), a content or offer change that made messages less recognisable or wanted (rebranding, new product category, changed From: name), a frequency increase that exceeded recipients' tolerance (same audience, more sends per week than previous pattern), or an authentication change that broke DKIM alignment (infrastructure migration, new sending service, DNS configuration change). Each diagnosis has a specific remediation: segment suppression, content adjustment, frequency reduction, or authentication fix.

The systematic diagnosis — timeline identification, cause isolation, specific campaign correlation — takes 30–60 minutes when the operational database and Postmaster Tools data are both available. Without the operational database (no accounting log pipeline), the campaign correlation step is missing and the diagnosis must rely on approximate timing and manual investigation. This is one of the most concrete operational costs of not building the logging infrastructure described in the logging architecture note: deliverability problems take longer to diagnose, and longer diagnosis means more accumulated reputation damage before remediation begins.

Domain reputation is the primary Gmail deliverability lever. Everything else — IP management, content optimisation, authentication precision, list hygiene — ultimately serves to maintain or improve the domain reputation signal that determines whether messages reach the inbox. Mastering the mechanics of how that signal is built, measured, maintained, and recovered is the highest-value deliverability knowledge available for Gmail-heavy sending programmes. The investment in this understanding pays compound returns over the lifetime of every domain that implements it.

The Sender Score Parallel and Why Domain Reputation Replaced It

Before Gmail's domain reputation system became the primary deliverability signal, IP-level reputation scores — measured by services like Sender Score — were the dominant deliverability indicators. Sender Score assigns a numeric score (0–100) to sending IP addresses based on complaint rates, bounce rates, and other signals collected from a panel of ISPs. A Sender Score above 90 was generally associated with good deliverability; below 70 indicated significant deliverability risk.

Gmail's domain reputation system differs from Sender Score in three important ways. First, it is domain-based, not IP-based, reflecting the operational reality that sophisticated senders rotate IPs while maintaining consistent sending domains. Second, it is proprietary to Gmail — the signals, weights, and thresholds are determined by Google and not disclosed in detail. Third, it directly affects Gmail delivery in a documented way, unlike Sender Score which was an aggregate indicator rather than a Gmail-specific signal.

Sender Score and similar IP reputation services remain useful for monitoring IP health across a broad panel of ISPs, but they are secondary to Gmail Postmaster Tools domain reputation for programmes where Gmail is the primary ISP destination. The operational priority is clear: Postmaster Tools domain reputation first, IP reputation services (Sender Score, SNDS, Postmaster Tools IP reputation) as supporting context. A programme that optimises Sender Score while neglecting Postmaster Tools domain reputation has its attention on the less important signal for most consumer email programmes.

Understanding the mechanics of Gmail domain reputation — how it is built, what moves it, how to measure it, and how to recover it — is the foundational competency for email infrastructure management at scale. Every other deliverability practice either contributes to or detracts from the domain reputation signal that determines whether email reaches Gmail inboxes. Making that signal consistently excellent, through the operational practices documented in this note, is the durable competitive advantage that well-managed email infrastructure builds over time.