Google says Gmail blocks more than 99.9% of spam, phishing, and malware, close to 15 billion unwanted emails a day. And yet the “quick question” from a sales rep you’ve never met sat at the top of your inbox this morning, above mail from actual colleagues. The filter works exactly as built, and this message was built to pass it.
Why Gmail won’t call a sales pitch spam
Gmail’s filter earns its 99.9% against mass mail, and mass mail has fingerprints: thousands of near-identical messages, links to known-bad domains, forged senders, spikes in user reports. Roughly 45% of everything sent worldwide fits that profile (Kaspersky measured spam at 44.99% of global email traffic in 2025), and the filter feeds on it.
Now audit a competent cold email against that checklist. It comes from a real person with a name and a LinkedIn profile, at a real company domain with SPF, DKIM, and DMARC configured. It’s addressed to you alone. The first line mentions your company or your job title, so no two copies of the campaign are identical. Volume per mailbox stays deliberately low. Nothing about it pattern-matches to the junk Gmail was trained on, and the sender’s job depends on keeping it that way.
Google’s own sender rules sharpen the irony. Since 2024, anyone sending more than 5,000 messages a day to Gmail addresses must authenticate their mail, offer one-click unsubscribe on marketing and subscribed mail, and keep user-reported spam under 0.3%. Those rules crushed a lot of sloppy bulk mail, and the cold outreach industry adapted instantly: sending tools now spread campaigns across many mailboxes and many domains precisely so no single sender ever resembles a bulk sender.
One more reason Gmail is gentle here: the mail is legal. The CAN-SPAM Act doesn’t require your consent before a stranger emails you a pitch. It requires truthful headers, a working opt-out honored within 10 business days, and it backs that with FTC penalties of up to $53,088 per email, but the sending itself is permitted. Gmail can’t treat a lawful, authenticated, individually addressed message as abuse without burying real mail from strangers too: the customer inquiry, the reporter, the person who found your talk online. So it doesn’t. Our guide to stopping spam in Gmail covers what the filter does catch; this page is about the lane it leaves open.
The pipeline behind the “quick question”
The message reads like one person typing to another. What produces it is a pipeline, and the vendors advertise each stage openly.
Your address gets found first: Apollo, one of the biggest prospecting databases, advertises 230M+ verified contacts on its own product pages. A sending platform runs the campaign; Instantly, one of the larger ones, lists plans from $47/month with unlimited email accounts and unlimited warm-up on every tier, scaling to a $358/month plan rated for 500,000 emails a month. “Unlimited accounts” is the load-bearing phrase: campaigns fan out across dozens of mailboxes on fresh domains, each staying under the radar individually. A Google Workspace account may send up to 2,000 messages a day, and outreach tooling keeps each inbox far below any threshold that would attract attention.
Backlinko and Pitchbox analyzed 12 million outreach emails and found 8.5% got any response at all; Woodpecker’s separate analysis of 20 million sales emails puts the 2026 average reply rate at 3.43%. The two studies measure different things, any response versus a counted reply, years apart, but they agree on the shape: replies are the exception. At those rates, a seller doesn’t need you to be interested. They need volume, which the tools make nearly free, and persistence, which the same Backlinko data rewards: a single follow-up lifts replies by 65.8%. Every “just bumping this to the top of your inbox” you’ve ever received was that percentage doing its work.
The manual toolkit, and where each piece gives out
Gmail’s toolkit for sales mail breaks into six moves, each effective against a specific shape of sender:
| Method | Stops | Outrun by |
|---|---|---|
| Block the sender | That one address, permanently | The same company’s next rep, or the same rep’s next domain |
| Filter the domain | Everyone at one company | Tomorrow’s pitch from a company you’ve never seen |
| Filter on phrases (“quick call”, “circle back”) | The clumsiest campaigns | Personalized copy, plus your own colleagues using the same phrases |
| Reply asking to be removed | Usually that one sequence | Confirms a live address; the next campaign starts fresh |
| Unsubscribe, when offered | Compliant senders’ future campaigns | Pitches styled as personal notes, which rarely carry a real opt-out |
| Report spam | Trains Gmail on true mass mail | One-to-one mail from the next clean domain, which carries none of the reported fingerprints |
Blocking a repeat sender and filtering a whole domain both work as far as they go; our guide to blocking emails in Gmail covers every variant, and unsubscribing safely once you can see who’s actually sending is covered in our guide to that too.
Every row in that table acts on an identity: an address, a domain, a literal string. The industry you’re up against rotates identities as a core competence, with databases holding hundreds of millions of contacts on one side and platforms advertising unlimited sending accounts on the other. You’re playing whack-a-mole where new moles cost the attacker nothing and each mallet swing costs you a trip to Settings.
Filtering on intent instead of identity
The durable fix has to classify the message, not the sender: not “is this address on my list” but “is this an unsolicited pitch.” That judgment doesn’t care which mailbox or domain the campaign rotated to this week, so rotation stops being a countermeasure.
Sieve is ours to sell, and it’s built for exactly this job. Read the rest of this section as the pitch it is.
Sieve sits behind Gmail’s own filter, Gmail-only by design, and classifies what that filter admits. Cold outreach, sales sequences, and recruiter pitches get a Solicitor label and are archived out of the inbox; phishing gets Malicious, with links screened via Google Web Risk. Archived means archived: nothing is deleted, everything stays searchable under its label, and dragging a mislabeled message back to the inbox is itself the correction, recorded automatically to improve future classifications.
Recruiter mail is the one category of cold outreach you might want someday, which makes blocking (permanent, per-address) exactly the wrong shape. Sieve takes instructions in plain English: tell it you’re job hunting this quarter and recruiter mail walks back into the inbox; tell it you’re done and the label returns.
Replying to any sender grants them permanent trust, which is the system working as intended: you replied, so it’s a conversation now, and one more reason not to answer pitches you never want to see again. Mail from people you know is never read by the AI at all; known senders skip classification entirely. Storage is metadata-only, plus the encrypted subject line of flagged mail held for 30 days, and none of it trains models. The Privacy Policy has the complete accounting.
For the minority of sales mail that does carry a legitimate opt-out, Sieve’s Smart Unsubscribe scans your last 90 days of message headers, sends a real RFC 8058 one-click request where the sender supports it, and blocks and files the sender either way, so the request being ignored costs you nothing.
Sieve costs $10/month for Basic, or $20/month for Pro with a second inbox included, 15% off yearly, each with a 7-day free trial. If you’re weighing it against the broader field, our spam-filter roundup puts it next to six alternatives, reasons to pick them included.
If a pitch a week is your whole problem
Then don’t pay anyone. Block the repeat offenders, filter the one relentless agency, and move on: the manual toolkit is genuinely enough at low volume, and it’s free. Paid tools earn their keep at the point where new senders arrive faster than you can name them.
And if what’s drowning you isn’t cold outreach, a different tool fits better. For newsletters you once signed up for, Leave Me Alone is built to unsubscribe you in bulk rather than block anything. For years of backlog clutter, Clean Email and Mailstrom are the bulk cleaners, compared head-to-head in our cleanup-tool comparison. And if you want your whole inbox re-sorted by priority instead of cleaned out, that’s SaneBox’s actual pitch. The rest of our guides take each of those jobs one at a time.
The pitch from a stranger, polite, personalized, authenticated, and utterly unwanted, keeps landing because Gmail’s filter was built to catch mass mail, not this.
Sources
- Google: New Gmail protections for a safer, less spammy inbox (99.9% blocked; ~15 billion unwanted emails/day)
- Google, Email sender guidelines (5,000/day bulk-sender threshold, one-click unsubscribe, 0.3% spam-rate rule)
- Google Workspace knowledge base: Gmail sending limits (2,000 messages/day per account)
- Kaspersky Securelist: Spam and phishing report for 2025 (spam = 44.99% of global email traffic)
- FTC, CAN-SPAM Act: A Compliance Guide for Business (no prior-consent requirement, 10-business-day opt-out, penalties up to $53,088 per email)
- Backlinko × Pitchbox: We Analyzed 12 Million Outreach Emails (8.5% response rate; one follow-up lifts replies 65.8%)
- Woodpecker: Cold Email Statistics Based on Sending Over 20M Cold Emails (3.43% average reply rate, 2026)
- Apollo.io product page: B2B contact search (230M+ verified contacts, as advertised July 2026)
- Instantly.ai pricing page (plans from $47/mo, unlimited email accounts and warm-up, 500,000 emails/mo top tier, as of July 2026)
- RFC 8058, Signaling One-Click Functionality for List Email Headers (IETF, 2017)