Graymail vs Spam: How Mailbox Providers Actually Sort Cold Email
A cold email to a named decision-maker is not automatically spam, even though the recipient never opted in — mailbox providers run a middle category for exactly this case, and understanding it changes what you optimize for. Get the classification logic backwards and you chase the wrong fixes: rewriting subject lines when the real problem is send volume, or worrying about spam-trigger words when the real problem is an unauthenticated domain. This is how Gmail, Outlook and other major providers actually decide where unsolicited but legitimate B2B email lands.
- Graymail is a distinct classification tier between inbox and spam — legitimate, unsolicited mail that isn't fraudulent or bulk-blasted.
- Inbox placement is decided mostly by engagement and volume signals, not by clever subject-line wording.
- Address-based, low-volume, one-to-one cold email naturally sits in the graymail lane; templated blasts push it toward spam.
- Spam complaints and delete-without-reading are the two engagement signals that do the most damage to sender reputation.
- Open-rate data is too distorted by privacy protections to use as a primary inbox-placement signal — watch replies and complaints instead.
Graymail: the classification tier between inbox and junk
Mailbox providers sort incoming mail into more than two buckets. There is the primary inbox, a lower-priority zone that Gmail exposes as the Promotions or Updates tab and that other providers just quietly deprioritize, and the spam folder. Graymail is the industry term for mail that lives in that middle zone: it comes from a real sender with a real business behind it, it is not phishing or malware, but the recipient did not ask for it and has not built up enough engagement history to earn default inbox placement. A cold email to a named decision-maker at a company you researched is graymail by definition — solicited it is not, spam it also is not, provided the sender behaves like a legitimate business.
The distinction matters because the fixes for stuck-in-graymail and flagged-as-spam are almost opposite. Graymail placement responds to engagement over time — replies, manual moves to inbox, added-to-contacts signals — and to keeping content and formatting plain enough not to trip content filters. Spam classification responds to reputation damage: complaint rates, spam-trap hits, blocklist entries, authentication failures. Treating a graymail problem like a spam problem, by rewriting copy or changing send times, wastes effort. Treating a spam problem like a graymail problem, by waiting for engagement to build, lets a burned domain degrade further.
None of the major providers publish an explicit graymail label a sender can check. What you actually observe is proxy behavior: mail landing in Promotions or Updates rather than the focused inbox, delivery with no bounce but also no visible complaint, and, over repeated sends from the same domain, a slow drift toward or away from the primary inbox. Cold email classification is a moving average, not a single verdict per message.
The three signal groups that decide where a message lands
Every major mailbox provider runs incoming mail through the same three signal groups, weighted differently by provider. Knowing the order they get evaluated in tells you where to spend effort.
- Authentication and network reputation: SPF, DKIM and DMARC alignment; the sending IP's and domain's history; whether the domain is new or has years of clean sending behind it. Fail this layer and content quality is irrelevant — the message can be rejected before any spam filter even reads it.
- Content signals: link density, image-to-text ratio, HTML-template fingerprints from mass-mail platforms, known spam-trigger phrasing, and whether the message reads as plain human correspondence or a marketing template.
- Engagement signals, aggregated across the sending domain and IP over time: open rate where it is measurable, reply rate, manual moves from Promotions to inbox, deletes without opening, and — the single heaviest-weighted signal — spam complaints, the 'report spam' click.
What keeps a cold B2B email in the graymail lane
The senders who consistently land in the focused inbox rather than Promotions or Spam share a pattern, and it maps closely onto how address-based outreach is supposed to work anyway: one named recipient per send, researched content that reads as correspondence rather than a campaign, and volume low enough that no engagement-based filter has reason to flag the domain.
- One email per send, addressed to a named person — not a merge-tag blast to a list, even if each copy is technically personalized.
- Plain-text or lightly formatted HTML with no tracking-pixel-heavy templates, stacked images, or unsubscribe-footer boilerplate lifted from an ESP.
- A reply-to address that is a real, monitored inbox — providers weight two-way mail flow, sent then replied-to, heavily in favor of the sender.
- A domain and IP with sending history and volume consistent with a person, not a campaign — dozens of sends a day from a given mailbox, not thousands.
- Content grounded in something specific to the recipient's company, which naturally avoids the generic phrasing content filters are tuned against.
This is also the practical argument against bulk ESP-style sending disguised as personalized cold email: the moment volume and formatting look like a campaign, providers apply campaign-grade scrutiny built to catch exactly the promotional-but-unsolicited pattern that pushes mail out of the primary inbox.
What actually pushes graymail into spam
Graymail turns into spam through accumulated reputation damage, not through any single email. The triggers are specific and mostly avoidable.
- Spam complaint rate above roughly 0.1% of sends — this is the fastest way to burn a sending domain; a few dozen report-spam clicks on a small send can do more damage than thousands of ignored-but-not-reported emails.
- Volume spikes from a previously low-volume identity — a mailbox that sent twenty emails a day suddenly sending five hundred reads as either compromised or campaign tooling.
- Identical or near-identical content sent to many recipients from the same domain in a short window — fingerprinting catches this even when merge tags vary the name.
- Hitting a spam trap: an address scraped from an old list, a purged database, or a honeypot address that never opts into anything — a single trap hit can flag a whole sending IP.
- Missing or misaligned DMARC, SPF or DKIM, especially after a domain or ESP migration — an easy miss that silently reclassifies mail that used to land fine.
- Aggressive follow-up cadence — three or more unanswered follow-ups on the same thread within days reads as automation to both the recipient and the filter.
Note what is absent from this list: subject-line wording, the word 'free', or other folklore spam-trigger phrases. Modern filters are engagement-and-reputation-driven; content-word blocklists are a legacy signal that still exists but rarely decides classification on its own for a well-authenticated, low-volume sender.
Benchmark numbers for healthy inbox placement
A few practitioner ranges are useful as sanity checks, not as targets to chase mechanically.
- Spam complaint rate: keep it under 0.1% of sends; anything above roughly 0.3% typically triggers provider-side throttling or blocklisting within days.
- Bounce rate: under 2-3% for a properly verified list; sustained bounce rates above 5% read as list-quality problems and drag reputation down regardless of content.
- Reply rate: a healthy cold B2B email campaign to a well-qualified ICP list runs 3-8% replies; below 1% usually means targeting or relevance is off, not deliverability.
- Domain warm-up: a new sending domain or mailbox typically needs 2-4 weeks of gradually increasing, engaged sending before it earns default inbox placement for cold mail.
Open rate deliberately is not on this list as a primary metric. Apple Mail Privacy Protection and similar image-proxy prefetching in Gmail trigger opens automatically on a meaningful share of mail regardless of whether a human read it, which distorts open-rate data enough that using it to judge inbox placement will mislead you. Replies and complaints are harder to fake and far more informative for cold email classification.
How LDM keeps cold outreach in the graymail lane
The platform's default posture assumes every cold email is competing for graymail classification, not inbox-by-default status, and builds the sending model around that instead of around bulk-marketing habits carried over from ESP tools.
- ICP-filtered recipient lists, not scraped or purchased ones — every send goes to a named decision-maker matched against defined criteria, which keeps volume naturally low per domain.
- Per-mailbox daily send caps well below what a marketing platform would push, mirroring how a person actually emails.
- Domain and mailbox warm-up scheduling before a new sending identity goes into full campaign use.
- Monitoring built around reply rate and complaint signals rather than open rate, for the accuracy reasons above.
- One-to-one message construction — researched, addressed to one person, formatted as correspondence — rather than a merge-tag template dressed up as personalization.
None of this guarantees every message lands in the primary inbox — graymail classification also depends on the recipient's own filter settings and history with senders in your category. But it keeps the controllable variables, authentication, volume, formatting and engagement pattern, pointed at graymail rather than spam, which is the realistic target for cold B2B email and the one address-based outreach is built to hit.
FAQ
Is cold email always classified as spam?
No. Mailbox providers distinguish unsolicited email from spam based on sender legitimacy, authentication and content pattern — a well-authenticated, low-volume email addressed to one person is graymail, not spam, even though the recipient never opted in. Spam classification requires additional negative signals: high complaint rates, spam-trap hits, or content patterns that match known abuse.
What is the difference between graymail and the Promotions tab?
They overlap but are not identical. Promotions and Updates is Gmail's specific UI implementation of deprioritized-but-legitimate mail; graymail is the broader classification concept other providers apply without a visible tab. A cold email landing in Promotions is a visible example of graymail classification in action.
Does using someone's first name in the subject line help avoid spam filters?
Not meaningfully on its own. Personalization tokens do not move authentication or reputation signals, which is what filters weight most heavily. What actually helps is the underlying pattern personalization should reflect — one recipient per send, relevant content, a real reply-to address — not the surface-level insertion of a name.
How long does it take a new sending domain to stop looking like graymail?
Typically 2-4 weeks of gradually increasing, engaged sending, assuming authentication is correctly configured from day one. Domains that skip warm-up and start at full cold-outreach volume immediately are far more likely to get stuck in graymail or slip into spam, because the provider has no clean history to weigh against the sudden volume.
Should I worry about specific spam trigger words?
Less than most advice suggests. Content-word blocklists are a legacy layer that still exists but rarely decides classification for a well-authenticated, low-volume sender — engagement and reputation signals dominate. Worry more about spam complaint rate and list quality than about avoiding words like 'free' or 'guarantee'.
Can a single bad campaign push a domain from graymail into spam permanently?
It can do serious damage, especially a spam-trap hit or a complaint spike, but it is rarely permanent. Recovery means stopping sends from the affected identity, fixing the root cause in list quality, volume or authentication, and rebuilding reputation with a slow, low-volume warm-up — the same process as onboarding a new domain, just starting from a deficit instead of neutral.
Want to apply this to your outreach?
We will map it to your segment and product — before any work starts.
Talk to us