What Is Email Deliverability and Why It Makes or Breaks Outbound

Most outbound teams treat deliverability as a setup task they finish once and never look at again, which is why the reply rate quietly dies in month 3 and nobody can explain it. We run outbound for 50+ B2B companies and have sent over 8 million cold emails this year, and when a campaign goes sideways, placement moves before copy does, every time. Below, the full measurement stack, the thresholds Google, Yahoo, and Microsoft actually enforce, and the order you diagnose a drop in.

What Is Email Deliverability?

Think of it as the last few inches of the journey. Your email can leave your server, pass through the receiving mail provider, and still get filed into a folder the recipient never opens. That filing decision is deliverability. Gmail, Outlook, and Yahoo make it in milliseconds, using signals they collected about you as a sender long before this specific message showed up.

For outbound, this is the whole game. You can write the sharpest, most relevant email in your market, and if it lands in spam your reply rate is zero. Deliverability sits underneath everything else. It is the foundation the copy, the targeting, and the list all stand on, and when it is broken nothing built on top of it counts.

It is also the part of outbound that fails silently. A bad subject line shows up as a low open rate you can see. Bad placement shows up as nothing at all: the sends go out, the dashboard turns green, and the replies just never arrive. That is why teams rewrite copy for 3 months while the actual problem sits in DNS.

Email Deliverability

The percentage of sent email that arrives in the recipient's primary inbox rather than spam or promotions. It is an inbox placement measure driven by sender reputation, authentication, list quality, and engagement. This is the number that predicts replies.

Inbox Placement Rate

The specific metric that quantifies deliverability. If 100 emails are accepted and 83 land in the primary inbox, the inbox placement rate is 83 percent. The other 17 are sitting in spam or promotions where almost nobody looks.

Sender Reputation

The running score each mailbox provider assigns to your sending domain and IP based on past behavior: complaint rate, bounce rate, authentication results, spam trap hits, and recipient engagement. It is per provider, not global, so Gmail and Microsoft can rate the same sender very differently.

What Is the Difference Between Delivery, Deliverability, and Inbox Placement?

These three get used interchangeably and the confusion costs real revenue. Delivery means the receiving server accepted your message and it did not bounce. Deliverability means that accepted message reached a folder a human actually reads. Inbox placement rate is the number that measures it.

Most sending tools report delivery, because delivery looks good and is trivial to measure. They show a 98 percent delivered figure and you assume the campaign is healthy. But delivered is not seen. The provider accepted the message, routed it to spam, and the dashboard still scored it as a win. This is the single most common reason a campaign with excellent delivery produces no replies.

Read that table once more and notice the middle row. Nobody hands you deliverability. There is no dashboard anywhere that reports the true share of your mail sitting in a real prospect's primary inbox, because the only party who knows is the mailbox provider and they do not publish it per sender. Everything else on that list is a proxy. Getting good at deliverability is mostly getting good at triangulating from proxies, which is why the measurement section below is the longest one on this page.

Why Does Deliverability Decide Whether Outbound Works?

Outbound is a volume and relevance game played against a hard ceiling, and the ceiling is the inbox. Every metric downstream is multiplied by placement. If 17 percent of your sends never reach the inbox, you lost 17 percent of your booked conversations before a single prospect read a word. According to Validity's 2026 email deliverability benchmark research, the global average B2B inbox placement rate sits near 83 percent, and that is the baseline, not the bad case.

The average also hides the part that matters most to B2B senders. Placement is not uniform across providers. Gmail averages close to 90 percent while Microsoft mailboxes, which is where a large share of enterprise buyers sit, run closer to 77 percent. If your ideal customer profile skews toward Outlook heavy industries like manufacturing, logistics, finance, or healthcare, your blended number can look fine while the segment you actually care about is being filtered. Read placement per provider or you are reading a number that describes nobody.

It gets worse when reputation slips. A new domain with no history, a list full of dead addresses, or a message that draws complaints can push placement to 50 percent or lower. At that point half your sends are invisible, reply rate halves, and the campaign looks broken even though the copy is fine. The owner blames the writing, rewrites the email, and never touches the real problem. We wrote the full teardown of that failure pattern in when cold email stops working.

Deliverability also compounds in both directions. Providers watch how recipients treat your mail. Land in the inbox and get opened and replied to, and reputation improves so the next batch lands better. Land in spam and get ignored, and reputation decays so the next batch lands worse. That feedback loop is why a small early problem becomes a dead domain in 6 weeks if you ignore it, and why the recovery is always slower than the collapse.

What Actually Decides Where Your Email Lands?

Providers do not publish their filtering rules, but the signals they weigh are well documented in their own sender guidance. Almost every placement problem traces back to one of six levers being neglected.

• Authentication. SPF, DKIM, and DMARC are the DNS records that prove the mail is really from you. Google's sender guidelines require all three from senders above 5,000 messages a day, with the From domain aligned to either the SPF or the DKIM domain. Microsoft applies the same 5,000 a day bar for Outlook mailboxes and rejects unauthenticated bulk mail outright. Full setup in what is SPF, DKIM, and DMARC and email DNS records explained.
• Sender reputation. Every domain and IP carries a score built from past behavior. New domains start with nothing and have to earn trust slowly, which is the entire reason warmup exists. See what is domain reputation.
• List quality. Dead addresses bounce, and bounces are the fastest way to wreck a young domain. Verify before sending, every time. See email verification and why it matters and cold email bounce rates, causes and fixes.
• Complaints and spam traps. A complaint is a recipient telling the provider your mail is unwanted. A spam trap is the provider finding out on its own, and trap hits are what get you onto a blocklist like Spamhaus rather than merely filtered. See cold email blacklist monitoring.
• Engagement. Opens, replies, and the absence of complaints tell providers a human wants this. Low engagement says the opposite. This is also why reply rate beats open rate as a health signal, covered in cold email reply rate benchmarks.
• Volume and pacing. A thousand identical messages from one address in an hour reads as a machine. A steady modest number from each of many mailboxes reads as a person. The pattern matters as much as the content. See how many cold emails to send per day and what is a mailbox rotation.

One thread runs through all six. Providers are trying to answer a single question: does a human want this email. Every lever is a different way of proving the answer is yes. Stop looking for tricks, start looking like a legitimate sender, and the levers organize themselves.

How Do You Measure Email Deliverability in 2026?

This is where most teams are flying blind, and it got harder in the last year rather than easier. Nobody reports your true inbox placement, so you build the number out of four independent sources that each see a different slice. Use one alone and you will draw the wrong conclusion.

The Postmaster Tools change matters more than anyone noticed. Google retired the domain reputation and IP reputation panels when senders moved to Postmaster Tools v2 in late 2025, on the reasoning that a High, Medium, Low, or Bad label was not actionable. Spam rate, authentication results, and the compliance status panel are what remains. If your team still describes a domain as having Good reputation, it is reading a screen that no longer exists. The modern equivalent is a spam rate held under 0.10 percent plus a passing compliance check.

Seed tests are directional, not a score. A seed list is a set of controlled addresses across providers that reports which folder your message landed in, and Twilio's guidance on seed list testing is blunt about the limitation: seed addresses are not real subscribers. They have no engagement history, they do not behave like people, and consumer providers have gotten good at recognizing them. Use seeds to catch obvious problems and to track relative movement between weeks. Never report a seed result as your deliverability. We cover the full method and the traps in inbox placement tests.

Your reply rate by sending domain is the most underrated instrument you own. It is the only number in the stack generated by real prospects. Segment it by domain and by mailbox, not just by campaign, and a deliverability problem announces itself as a shape: one domain flat while its siblings hold, or every mailbox on one provider dropping together. That shape tells you where to look before any dashboard does. The monitoring routine is in cold email deliverability monitoring, and the wider reporting layer in how to track cold email campaign performance.

What Do the Numbers Have to Be?

The thresholds below are not opinions. They come from the providers' own published sender rules, which means crossing them triggers an automatic penalty rather than a negotiation.

Google tells bulk senders to keep spam rates reported in Postmaster Tools below 0.30 percent and recommends staying under 0.10 percent, to authenticate with SPF and DKIM plus aligned DMARC, and to support one-click unsubscribe per RFC 8058 on marketing and subscribed mail. Yahoo's sender best practices mirror the complaint threshold and add a DKIM key length minimum and valid reverse DNS on every sending IP. Microsoft applied the same 5,000 a day bar to Outlook mailboxes, rejecting unauthenticated high volume mail rather than filtering it. Mailgun's research on the Google and Yahoo rollout found roughly 80 percent of senders who reacted changed their authentication setup, which tells you where the industry's weak point was.

Two notes on reading that table. The reply rate row is a health signal, not a deliverability metric, but it is the one that moves first when placement slips, which is why it belongs here. And the DMARC row is about policy strength, not presence: a DMARC record sitting at p=none passes the bulk sender check while enforcing nothing, so it buys compliance without buying protection. The ladder from none to reject is walked step by step in email DNS records explained.

How Do You Diagnose a Deliverability Drop?

When replies fall off, the instinct is to rewrite the email. That is almost always the last thing to check, not the first. Work this order and you will find the cause in under an hour instead of spending a month on copy that was never broken.

1. Check whether it is one domain or all of them. Pull reply rate by sending domain for the last 14 days. One domain down while its siblings hold is a domain level problem: reputation, a listing, a DNS change. Every domain down at once is a list, copy, or targeting problem, or a provider side change that hit the whole fleet.
2. Check whether it is one provider or all of them. Split replies by recipient domain, Gmail versus Microsoft versus everything else. Microsoft only usually means an SNDS or IP issue. Gmail only usually means complaints or authentication alignment.
3. Read the spam complaint rate. Open Postmaster Tools before anything else. If the spam rate crossed 0.10 percent, you found it and nothing else on this list matters until it comes down.
4. Read the bounce rate. A bounce spike means a list problem: a bad source, an expired verification, or a segment that was never verified. Bounces above 2 percent damage reputation while you investigate, so pause that segment first and diagnose second.
5. Check blocklists. Run every sending domain and IP against Spamhaus and the major lists. A listing explains a cliff edge drop that nothing else can.
6. Verify authentication still passes. DNS changes break SPF and DKIM silently, and a hosting migration or a new tool added to the stack is the usual culprit. Re-check alignment, not just presence.
7. Check volume and ramp. Did somebody raise sends per mailbox, add mailboxes without warming them, or restart a paused campaign at full volume. Sudden volume changes read as compromise to a filter.
8. Only now, look at the copy and the list. If the first 7 checks are clean, the problem is upstream of sending: the wrong people, or the wrong message to the right people. That is a targeting question, and it starts at how to define an ICP for cold email.

The reason this order works is that it moves from cheapest and most likely to most expensive and least likely. Steps 1 through 3 take about 10 minutes and catch most incidents. Rewriting copy takes a week and catches almost none, which is exactly backwards from how most teams spend their time. If the domain is already burned past repair, the recovery path is in how to recover a burned domain.

How Do You Build So It Does Not Break?

Protecting deliverability is infrastructure and discipline, not clever writing. The setup that holds at volume is the same one every serious outbound program runs, and it starts with keeping cold email off your primary domain entirely.

1. Send from secondary domains. Buy separate domains for outreach so a rough sending week never touches the mail your company runs on. Structure and naming in what is a secondary domain and how to set up email domains for outbound.
2. Authenticate before the first send. SPF, DKIM, and DMARC on every sending domain, with the From domain aligned. Start DMARC at p=none with reporting on, read the reports for 2 weeks, then move to quarantine.
3. Warm every mailbox. Run each new mailbox through 3 to 4 weeks of warmup, ramping volume gradually so the provider watches a normal sender build a normal history. Method in email warmup explained and how to warm up a new email domain.
4. Verify every address. Run lists through a verifier before they enter a campaign. This one step prevents most reputation damage we see, and it costs a fraction of what a burned domain costs to replace.
5. Cap volume per mailbox. Keep each mailbox near 30 to 50 sends a day and spread total volume across many mailboxes rather than pushing a few hard. Fleet sizing math in multi domain sending strategy.
6. Monitor weekly, not quarterly. Spam rate, bounce rate, blocklist status, and reply rate by domain. Put it on a recurring slot so it happens when nothing is wrong, because that is the only way you will notice the week something is.

When placement does drop, the recovery is to pull volume back, clean the list, fix the cause, and let warmup rebuild trust before scaling again. The instinct to push harder when results dip is exactly wrong: more volume on a damaged domain accelerates the decline. Google restores delivery help to bulk senders only after spam rates hold under 0.30 percent for 7 consecutive days, so patience is not a virtue here, it is a requirement written into the rules.

Where to Go Deeper on Each Layer

Deliverability is one word covering about a dozen separate disciplines. This page is the map. Each link below is the territory.

• Authentication: SPF, DKIM, and DMARC explained and the DNS records behind them.
• Infrastructure: cold email infrastructure setup, secondary domains, multi domain sending, mailbox rotation.
• Warmup and ramp: warmup explained, warming a new domain, recovering a burned one.
• List hygiene: email verification, bounce rate causes and fixes, list building from scratch.
• Filtering and spam: staying out of the spam folder, why cold email hits spam, blacklist monitoring.
• Measurement: inbox placement tests, deliverability monitoring, positive reply rate.
• Volume and pacing: how many to send, when to send, cadence design.
• Compliance: GDPR and CAN-SPAM for cold email.

The Practitioner Take on Deliverability

After 8 million sends, the lesson is blunt. Deliverability is the tax you pay before any other outbound work counts. Build a perfect list, write a perfect email, and if the infrastructure underneath is sloppy you will get a fraction of the result and blame the wrong thing. The teams that win are not the ones with the cleverest copy. They are the ones boring enough to authenticate every domain, warm every mailbox, verify every list, and read placement weekly when nothing appears to be wrong.

What changed in the last year is that the providers stopped handing out grades. Postmaster Tools took away the reputation label. Microsoft raised its bar for high volume senders. The measurement got harder while the enforcement got stricter, which punishes exactly the senders who were relying on a dashboard to tell them when to worry. The replacement is not a better dashboard, it is a habit: 4 sources, read weekly, triangulated.

There is a second order point here that most people miss. Deliverability is downstream of the ask. An email that makes a stranger feel useful draws replies and near zero complaints, and those two signals are exactly what providers reward. An email that makes a stranger feel sold to draws complaints, and complaints are what providers punish. Changing what you ask for changes the engagement profile, which changes placement. That is a large part of why an invite outperforms a pitch on the same list and the same infrastructure, which we broke down in why podcast invites beat pitches in cold email.

We run this stack for every client on the Reverse Outbound Engine: domains, mailboxes, warmup, verification, and monitoring all included, with the whole program backed by 30 recorded conversations with your ideal buyers in 90 days or your money back. The infrastructure is not the product. It is the price of admission to have a product at all.

Deliverability is not glamorous and never will be. It is plumbing. But the senders who treat the plumbing as the real product, and the copy as the thing that rides on top of it, are the ones still landing in 2026 while everyone else wonders why their reply rate quietly went to zero.