Why Your Emails Aren't Reaching the Inbox (and What Actually Fixes It)

Most deliverability problems trace back to one of three things: your domain hasn't earned enough trust with the mailbox provider, your list is telling that provider you're not wanted, or a technical record is misconfigured in a way that only shows up once volume crosses a threshold. Fixing deliverability means finding which of those three is actually happening to you, because the fix for each one is completely different, and applying the wrong one (say, switching ESPs when the real problem is complaint rate) just moves the same problem to a new IP.

Globally, about 87% of legitimate commercial email reaches the inbox, according to Validity's 2026 benchmark data, up from roughly 83% the year before. That still leaves a meaningful chunk of mail landing in spam folders or vanishing outright, and the gap between providers is wide: Gmail placement runs around 87%, while Outlook.com trails at closer to 76%. If you're only checking whether Gmail delivery looks fine, you're missing where the actual damage is happening.

"It's not delivering" usually means one of four different things

This is the part that trips people up first: "my email isn't delivering" is four different failure modes, and mixing them up wastes days of troubleshooting.

• Bounced. The receiving server rejected the message outright, hard (address doesn't exist) or soft (mailbox full, server temporarily down). This shows up in your ESP's bounce log.
• Filtered to spam. The message was accepted and delivered, just not to the primary inbox. Your ESP will often report this as "delivered" because technically it was, which is why bounce-rate dashboards can look healthy while open rates quietly die.
• Missing. Accepted, no bounce, never filed anywhere the recipient can find it. Some providers silently drop mail from senders they've decided aren't worth flagging as spam or delivering; it just disappears. This is the hardest failure to diagnose because there's no log entry telling you it happened.
• Blocked. The receiving server refused the connection entirely, often citing a blocklist or a reputation threshold, before the message content was even evaluated.

A domain with a real problem usually shows all four at once, in different proportions depending on the provider. Gmail tends to filter before it blocks; a strict corporate Microsoft 365 tenant tends to block before it filters. That's why a single "deliverability rate" number from your ESP dashboard is close to useless for diagnosis. You need provider-level breakdowns, which is what a seed-list testing tool or Google Postmaster Tools actually gives you.

Authentication is no longer optional at any real volume

Since February 2024, Gmail and Yahoo have enforced specific requirements on anyone sending what they define as bulk volume, roughly 5,000 messages a day to their users, combined across all sending domains. Below that threshold the rules are recommendations; above it, they're enforced with rejections and throttling. Mailgun's breakdown of the 2024 policy is the clearest summary of what actually changed, and it's worth reading in full if you send even close to that volume.

The detail that catches people off guard: DMARC at p=none satisfies the letter of the requirement but does nothing to stop spoofing, because p=none means "don't do anything if this check fails, just tell me about it." It gets you past the compliance checkbox without actually protecting your domain. Moving to p=quarantine or p=reject is what actually blocks lookalike-domain phishing that rides on your sending reputation, and Gmail and Yahoo do reward domains that enforce a real policy with better placement over time, not just compliance.

DMARC without SPF or DKIM already passing is a dead end, incidentally. DMARC doesn't authenticate anything on its own; it's a policy layer that tells receiving servers what to do when SPF or DKIM alignment fails. Get the underlying two working first.

The complaint-rate math that actually gets domains blocklisted

0.3% sounds tiny, and that's exactly why senders misjudge it. Google's own guidance is to treat 0.3% as the point of no return, and to actually target well under 0.1% for reliable placement. Here's what that looks like at real volume: if you send 100,000 emails and 300 people hit "report spam," you've hit the hard ceiling. At 100 complaints on that same volume, you're already in what Google calls the danger zone above 0.1%. That's not a big list problem; that's roughly one complaint per thousand recipients being enough to start throttling you.

The complaint that actually hurts isn't the person who unsubscribes. It's the person who doesn't bother finding the unsubscribe link and clicks "report spam" instead because it's faster. Every list I've cleaned up that had a real complaint problem had the same root cause: an unsubscribe flow buried two clicks deep, or a "manage preferences" page that required a login. One-click unsubscribe isn't just an RFC 8058 compliance checkbox for bulk senders anymore; it's the single highest-leverage fix for complaint rate, because it gives the annoyed recipient an easier exit than the report-spam button.

Reputation lives on the domain more than the IP now

Ten years ago, deliverability advice centered on IP warmup: get a dedicated IP, ramp volume slowly, protect that IP's reputation like it was the whole game. That's still true if you're sending tens of millions of messages a month on a dedicated IP, but for the large majority of senders on shared ESP infrastructure, domain reputation now matters more than the IP you happen to be routed through that day. Gmail's postmaster tools report reputation at the domain and From-address level specifically because shared IPs churn between thousands of tenants; scoring the IP alone would be nearly meaningless.

This is why switching ESPs doesn't fix a reputation problem, and it's the mistake I see most often. A team with a damaged domain moves providers, gets a temporary bump because the new IP pool hasn't seen their bad behavior yet, and watches placement collapse again within two to three sends once the new provider's domain-level signals catch up. The domain's history didn't reset; it just took a few days for the new mailbox providers to relearn it.

What the spec sheet won't tell you about list hygiene

A few things that don't show up in any deliverability checklist but matter in practice:

Recycled spam-trap addresses are the quiet killer. These are old addresses that mailbox providers reactivate specifically to catch senders with stale lists, and unlike bounce-triggering fake addresses, they accept the mail silently while dinging your reputation behind the scenes. If you haven't emailed a list in over six months, treat every address on it as a spam trap until proven otherwise, meaning a re-engagement send to a small segment first, not a blast to the whole file.

Image-heavy, text-light emails still get flagged by content filters more than people expect, not because filters are "detecting marketing" but because that ratio is a strong statistical signature of the phishing and low-effort spam templates filters are trained against. It's not a hard rule, but if your open rates on Gmail specifically underperform other providers on the exact same send, check whether your image-to-text ratio is the outlier.

Engagement history is now weighted more heavily than list size or send frequency by most major filters. A list of 20,000 people who open half your emails will outperform a list of 200,000 where 5% do, even though the second list looks like the bigger asset on a dashboard. Suppressing non-engagers isn't just good practice for your metrics; it directly protects the domain reputation of every email you send to people who are still paying attention.

Diagnosing an active problem

Start with Google Postmaster Tools if any real share of your list is on Gmail; it's free, requires a DNS TXT verification, and shows domain reputation, spam rate, and authentication success rate broken out day by day. Cross-check with a blocklist lookup (MXToolbox is the standard free option) to rule out the blunt-force cause before chasing subtler reputation issues. Then run a seed-list test, sending to a set of monitored addresses across Gmail, Outlook, Yahoo, and a couple of smaller providers, to see where placement actually differs. If Gmail looks fine but Outlook is filtering everything, you're not looking at a global reputation problem; you're looking at a Microsoft-specific one, and Microsoft's SNDS and JMRP tools are where that investigation continues.

How long recovery actually takes

This is the question everyone asks right after they fix the underlying cause, and the honest answer is that reputation repair is measured in sending cycles, not hours. A domain that's been suppressed for a genuine complaint-rate problem typically needs two to four weeks of clean sends, meaning low volume, highly engaged segments only, before providers meaningfully relax filtering. There's no support ticket that speeds this up; Gmail and Yahoo don't offer manual reputation reviews for marketing senders the way they do for confirmed false-positive blocklist cases. The fastest path back is consistency: same sending pattern, same authentication, engaged recipients only, until the pattern of good behavior outweighs the pattern of bad behavior in whatever scoring window the provider uses. Rushing volume back up during that window is the single most common way senders reset their own recovery clock.

Related Guides

SPF, DKIM & DMARC Setup for Business Email: The Complete 2026 Guide - the actual DNS TXT record steps for the authentication layer this article assumes you already have in place.

Domain Email Stopped Working After Website Migration - for when mail disappears right after a host or DNS change, the "missing" failure mode covered above but from the migration angle specifically.

How to Troubleshoot Gmail Errors in Google Workspace Admin - practical steps for the Gmail-specific diagnosis this article recommends via Postmaster Tools.

Microsoft 365 Email Not Working? 8 Real Causes & Fixes - for the Outlook/Microsoft placement gap the article flags as the weaker of the two major providers.