Cold email deliverability is the share of your cold emails that reach the primary inbox instead of spam, and it is decided mostly before anyone reads the copy: by your sending domains and authentication, your warm-up and daily volume, and the quality of your list. Gmail asks every sender to keep spam complaints under 0.3%. Reachly holds bounce rate under 2%, warms every new sending domain for 4 weeks, and takes a mailbox out of rotation as soon as its reply rate falls below 0.5%, its warm-up score drops below 95, or its bounces pass 2%.
Key Findings
- Deliverability usually breaks in the infrastructure, the data or the targeting long before the copy matters, so troubleshoot in that order.
- Send cold email only from dedicated secondary domains, with SPF, DKIM and DMARC finished before the first campaign.
- Warm new domains for 4 weeks before campaign traffic, keep daily volume per mailbox low, and add mailboxes and domains when you need more volume.
- Gmail wants spam complaints under 0.3%, and recommends under 0.10%. Gmail and Outlook both require SPF, DKIM and DMARC from domains sending more than 5,000 messages a day.
- Hold bounce rate under 2%. Above 5%, providers start to block.
- A burned domain is worth saving only when the cause was short-lived and is already fixed. Otherwise retire it and start a clean one.
- Keep about 30% of sending capacity in reserve so a tired mailbox can be rotated out without pausing the campaign.
If you need a copy layer once the infrastructure is sorted, the cold email response rate 2026 benchmarks guide covers what good looks like and how to lift yours.
Why are your cold emails really going to spam?
Most senders misdiagnose deliverability.
They think spam placement means the message is weak. Sometimes it is. More often, the primary issue is that mailbox providers do not trust the sender.
A lot of outbound advice is shallow. Change the subject line. Remove one link. Personalize the opener. Fine. Those things matter later. If your domain has no real reputation, your authentication is half-configured, and you are pushing mail through inboxes that were never warmed properly, the campaign is dead before the first prospect ignores it.
The real problem is fragile infrastructure
Cold email is an operating system, not a template.
You need a protected domain strategy, properly authenticated inboxes, strict volume control, clean data, and active monitoring. Miss one part and the rest starts failing in ways that look random. Open rates drop. Replies dry up. Bounce rates creep up. Then someone blames market saturation when the underlying issue is self-inflicted.
Your copy cannot rescue a sender reputation problem.
Most founders and SDR managers fall into this trap. They keep adjusting campaign copy because that feels easier than touching DNS, mailbox health, or list hygiene. That cycle explains why the same problem comes back every quarter.
Spam is a system failure
If you want inbox placement that holds up, stop treating deliverability like a one-time setup task.
Build the system so problems get blocked before they start. That means isolating your main domain, authenticating every sending domain, warming inboxes with discipline, verifying lists before launch, and watching reputation signals every day. Do that and deliverability becomes manageable. Skip it and every campaign becomes a gamble.
How should you set up domains and authentication for cold email?
A founder launches outbound from the company's main domain, books a few replies, then notices customer emails landing in spam two weeks later. That mistake is avoidable. Your cold email setup should contain risk, not spread it across the entire business.
Never send cold outreach from your primary domain.
Use separate sending domains built for outbound only. Pick close variants of your brand so the domain still looks legitimate to prospects, but keep your real company domain out of campaign traffic. If one sending domain gets flagged, your sales motions slow down. Your support, hiring, finance, and founder email stay untouched.
Separate domains by function
Treat outbound infrastructure like a portfolio, not a single sender identity.
Buy a small set of adjacent domains. Create outbound-only inboxes on each one. Keep campaign traffic, reply handling, and any testing inside that environment. Do not mix outbound with customer conversations or day-to-day company email. That separation is what gives you room to scale, pause, replace, and recover without taking down the rest of your email operation.
This is the part outbound teams rush. They buy domains, connect a sending tool, and assume the software solves trust. It does not. Sending platforms can automate volume and rotation. They cannot fix a bad domain plan or a broken DNS setup.
SPF, DKIM, and DMARC need to be finished before launch
Authentication is how mailbox providers verify that your mail is allowed to come from your domain.
- SPF lists the servers allowed to send on behalf of your domain.
- DKIM signs each message so receiving servers can verify it was not altered.
- DMARC tells receiving servers how to handle mail that fails those checks and gives you reporting on alignment and abuse problems.
If any one of these is missing or misconfigured, you create doubt at the exact point where providers decide whether to trust you. That doubt hurts placement before your copy, offer, or targeting gets judged.
A practical setup looks like this:
| Part | What to do | Why it matters |
|---|---|---|
| Primary domain | Keep it out of outbound campaigns | Protects normal business email |
| Sending domains | Use separate brand-adjacent domains | Contains risk to the outbound layer |
| Mailboxes | Create outbound-only inboxes on each domain | Prevents one identity from carrying all volume |
| SPF | Authorize your mail servers correctly | Stops unauthorized senders from claiming your domain |
| DKIM | Sign every message | Verifies message integrity |
| DMARC | Set policy and reporting | Catches alignment failures and spoofing |
What Gmail and Outlook require
Both providers publish their rules, so there is no need to guess. Since February 1, 2024, Google requires every sender to Gmail accounts to set up SPF or DKIM, keep valid forward and reverse DNS, send over TLS, and keep the spam rate shown in Postmaster Tools below 0.3%. Google recommends staying under 0.10%, which is the number to manage to, because 0.3% is where filtering starts rather than where it ends.
Domains sending more than 5,000 messages a day to Gmail must also run SPF and DKIM together, publish a DMARC record aligned with the From address, and support one-click unsubscribe on marketing mail. Outlook set the same bar for domains over 5,000 a day: SPF and DKIM must pass, and DMARC must be at least p=none and aligned. It started enforcing on May 5, 2025, and non-compliant mail is junked or rejected with a 550 5.7.515 error.
Treat the 5,000-a-day line as a floor for everyone, whatever your volume. A cold email program that splits its sending across many domains can sit under the threshold on every one of them, and setting up DMARC on each costs nothing. You can read your own Gmail spam rate for free in Google Postmaster Tools. The email deliverability guide sets the spam complaint and bounce limits to stay under once you can see that number.
Build for replacement, not perfection
Good outbound infrastructure assumes something will burn out.
A sending domain can lose placement. An inbox can drift. A mailbox provider can tighten filtering with no warning. If your whole program depends on one domain and a handful of inboxes, every issue becomes urgent. If you spread outbound across dedicated domains and properly configured inboxes, you can pause one lane, replace it, and keep the system running.
That is the goal here. You are not patching deliverability after it breaks. You are building an outbound base that survives normal wear without taking the company down with it.
The setup rules we use
We keep this simple and strict.
- Use brand-adjacent domains only for outbound. Never recycle the main company domain for campaign traffic.
- Create dedicated inboxes for sending. Outbound accounts exist for outbound only.
- Finish DNS setup before any campaign goes live. SPF, DKIM, and DMARC are configured and checked first.
- Keep roles separate. No cold outreach from founder, support, recruiting, or customer success inboxes.
- Plan for domain rotation. Have backup sending capacity ready before you need it.
Teams that follow this model avoid the usual pattern of short-lived results followed by a deliverability collapse. Teams that ignore it keep blaming copy, tooling, or market conditions for problems they created in DNS and domain strategy.
How do you warm up and scale cold email sending?
You buy a fresh domain, connect a few inboxes, push a campaign on day three, and watch reply rates die by lunch. That failure starts long before the copy gets judged. It starts with sending behavior that looks synthetic from the first week.
Warm-up is reputation building. Scaling is load management. Treat them as one system.
Mailbox providers do not grade your intent. They grade patterns. A new inbox that suddenly behaves like a sales robot gets filtered like one. The fix is simple. Build a believable history first, then increase volume in small steps that the inbox can absorb without changing its behavior overnight.
Warm-up should create normal mailbox behavior
A good warm-up period gives each inbox a baseline. It sends messages at uneven times, generates real thread activity, and avoids the sharp spikes that make a fresh account look automated. If you use a warm-up tool, keep the goal clear. You are not trying to inflate activity. You are trying to make the mailbox look stable, boring, and human.
Microsoft says sender reputation is shaped by user engagement and complaint patterns, not just raw send volume, in its guidance on email deliverability and sender reputation for Outlook. Google makes the same point in its Email sender guidelines, where bulk senders are told to keep spam rates low and maintain consistent sending behavior.
That is why warm-up fails for outbound teams that rush into campaigns. They focus on volume targets instead of behavior patterns.
Use a controlled ramp, not a dramatic launch
Keep the ramp plain.
- Start each inbox with low activity.
- Increase volume in small increments every few days.
- Keep warm-up traffic running while campaign volume comes online.
- Spread sends across multiple inboxes instead of forcing one account to carry the whole program.
- If you need more capacity, add more inboxes. Do not ask a single mailbox to become your entire outbound engine.
Keep each sending address well under what the tools allow. Reachly's 30,000-email month for one client runs on 180 mailboxes, which averages about 8 cold sends per mailbox per working day.
The operating rule we use:
- Warm each inbox before it touches prospecting traffic.
- Increase volume gradually, not in weekly jumps.
- Keep send timing irregular enough to avoid machine-like patterns.
- Hold back capacity on every domain so you can absorb issues without shutting down outreach.
- Scale by adding lanes, not by squeezing harder on the same lane.
That last point matters more than teams admit. Good outbound infrastructure is built for replacement. If one inbox slips, you should be able to pause it, swap in another, and keep campaigns running without a full reset.
Know when an inbox is ready to carry real volume
A warm inbox is not ready because a tool says it is. It is ready when it handles normal traffic without obvious stress signals.
Check the basics. Replies should come in without delays that feel abnormal. Placement should stay steady across major providers. Bounce behavior should stay clean. The inbox should tolerate a small increase in campaign traffic without a visible drop in engagement or a sudden rise in filtering.
Then scale carefully.
Do not stop warm-up and slam the account with production volume. Blend the two. Add live sends in controlled batches and watch what changes after each increase. If performance drops, pull volume back fast and isolate the inbox before the problem spreads across the domain.
If your targeting depends on shaky contact records, scaling gets harder because every inbox takes unnecessary hits. Clean prospect inputs matter as much as warm-up discipline, especially if you are enriching lists before launch with signal-based outbound workflows.
Nine times out of ten the problem is in the DNS, the warm-up, or the list. Fix those three, the copy you already have starts working again.
What actually breaks inboxes during scale
Outbound teams usually damage sender reputation the same way. They warm for a short stretch, see a few replies, then jump straight to campaign volume. Or they keep adding sends to the same inbox because it worked last week. Both are lazy operating decisions.
The safer model is boring by design. Multiple domains. Multiple dedicated inboxes. Controlled ramps. Ongoing warm activity. Spare capacity. Clear pause rules.
You do not fix cold email deliverability by waiting for placement to collapse and then hunting for a tool. You build a sending system that stays stable as volume grows.
How does list quality affect cold email deliverability?
You launch a clean campaign on a healthy inbox, then the reply rate dies and bounces stack up by day two. That usually is not a domain problem. It is a data problem.
Bad data poisons outbound faster than weak copy. One stale list can turn a stable sending setup into a reputation cleanup project. If you want deliverability that holds up at volume, treat list quality as infrastructure, not prep work.
Your list starts going bad the moment you export it
People change jobs. Companies shut down inboxes. Teams migrate domains. Shared spreadsheets keep circulating long after the records inside them are wrong.
ZeroBounce explains that email data decays quickly and recommends validating addresses before campaigns, not months in advance, in their analysis of email list decay and verification timing. Their guidance aligns with the rule good outbound teams already follow. Hard bounces should be removed fast, and old records should never be trusted without a fresh check.
Build a data gate before any inbox sees volume
Run outbound data through three filters.
Use better inputs. Start with sources that give you useful company and contact context, not just an email field. Job changes, firmographic filters, and recent signals matter because they help you avoid sending valid email to the wrong person. If your team is enriching lead records before launch, a structured signal-based outbound process cleans up most downstream deliverability issues.
Verify right before the send. Verification done too early creates false confidence. Mailgun recommends validating email addresses at the point of capture and again before sending because addresses can become invalid over time, in their email verification best practices. The standard sequence is simple: export, segment, verify, send. In that order.
Suppress permanently. Hard bounces, unsubscribes, spam complainers, and known bad-fit contacts should never re-enter rotation. HubSpot is clear on list hygiene in its email list management guidance: remove invalid and disengaged contacts to protect sender reputation and keep complaint risk down.
This process is boring. Good. Boring systems protect domains.
Valid emails still hurt you if the targeting is sloppy
A verified address is not the same thing as a good prospect.
Outbound teams often obsess over verification and ignore fit. Then they blast a large list of technically valid contacts who do not care, do not reply, and do not recognize the sender. That pattern tanks engagement, and low engagement gives mailbox providers another reason to filter you harder.
The fix is tighter segmentation before copy ever gets written. Filter by role, company profile, current pain, and recent trigger. Build smaller batches around specific offers. Relevance protects deliverability because it produces the positive signals your infrastructure needs to stay trusted.
Small targeted lists keep inboxes healthier than large lists full of weak-fit contacts.
What we cut from every list
Before a campaign goes live, remove these records:
- Role accounts like info@, support@, admin@, and hello@
- Catch-all or risky addresses your verifier cannot confirm with confidence
- Old records that have not been re-verified close to send time
- Previous hard bounces
- Unsubscribes and spam complainers
- Low-fit contacts added only to inflate volume
Outbound operators run into trouble when they keep weak records because they want a bigger top of funnel. Bigger volume does not help if the inputs are dirty. It just spreads reputation damage across more inboxes.
How do you monitor and troubleshoot cold email deliverability?
A campaign can look fine at 9 a.m. and be damaged by lunch.
That is why deliverability monitoring belongs inside your operating system, not as a cleanup task after results drop. If you wait until reply rates crash, you are already late. The job here is simple: catch trust loss early, isolate the cause fast, and stop bad sends before they spread reputation damage across the rest of your setup.
Watch the metrics that predict trouble
Open rate is a weak control metric. It is noisy, privacy-distorted, and easy to misread. Track the signals that tell you whether mailbox providers still accept your mail and whether your targeting still deserves inbox placement.
The core dashboard should include:
| Metric | What it tells you | What to do if it slips |
|---|---|---|
| Bounce rate | Verification, enrichment, or list freshness failed | Pause the campaign and inspect the affected segment |
| Spam complaints | Targeting or message angle is provoking negative reactions | Cut volume, tighten audience, review the offer |
| Provider-level opens | Gmail, Outlook, or others are filtering you differently | Compare inboxes by provider, inspect domain-level issues |
| Positive replies | Relevance is strong enough to support sender reputation | Narrow the segment, rewrite around a clearer pain point |
A specific Gmail-versus-Outlook example: if Gmail placement drops while Outlook holds steady, the likely culprit is DKIM alignment or a recent Google bulk-sender threshold change. Check your DMARC reports for that domain before touching anything else.
For pacing, follow-up control, and reply handling, your sequence logic matters too. Poor timing can create the same reputation problems as bad data. A clear signal-based outbound playbook keeps the system controlled as volume grows.
Troubleshoot in the right order
Do not start with copy.
Start with failure type, because each symptom points to a different break in the system. If you troubleshoot from the template first, you waste time and keep sending through a setup that may already be compromised.
Use this order:
- Bounce rate rises first. Assume bad records, stale verification, or a broken source.
- Spam complaints rise. Assume weak targeting, a bad offer match, or copy that sounds deceptive or irrelevant.
- One provider drops harder than the others. Check authentication alignment, recent volume jumps, and domain health for that provider.
- Replies disappear while delivery looks normal. Check segmentation, offer fit, and CTA friction.
That sequence keeps you from making the common outbound mistake of rewriting emails when the problem sits in list sourcing, infrastructure, or send behavior. Diagnose sender health first, then data quality, then targeting, then copy.
Match the symptom to the cause
Each symptom points to a different break. Read the pattern across providers before you change anything, because a drop at one provider and a drop everywhere have different causes.
| Symptom | Likely cause | First move |
|---|---|---|
| Bounces rising | Stale data or skipped verification | Pause the list and re-verify every contact before it sends again |
| Opens falling at one provider | Reputation problem with that provider only | Check recent volume to that provider, run placement tests, audit the domain |
| Opens falling everywhere | A domain-wide or setup problem | Pause the campaign and audit authentication and DNS end to end |
| Delivery normal, replies weak on broad lists | Relevance, not sending | Split the list by signal and rewrite the message for each segment |
| Spam complaints rising | Wrong audience or weak fit | Cut the failing segment and suppress complainers permanently |
The last two rows are where teams lose the most time, because they look like deliverability problems and are targeting problems. Track each signal segment separately, so a funding-triggered list and a hiring-triggered list never share one averaged reply rate. That is how Reachly runs campaigns: lists are segmented by signal in Clay, sent through Smartlead on dedicated domains, and coordinated with HeyReach on LinkedIn, so each segment gives a clean read.
Bounce handling needs rules, not judgment calls
Manual bounce cleanup is how bad addresses slip back into future sends. Build automatic suppression and keep it strict.
A healthy outbound system should treat these as default rules:
- Hard bounce once, suppress forever
- Unsubscribe once, suppress forever
- Complaint or hostile reply, exclude immediately
- Soft bounce patterns, review by provider and mailbox
- Old segments, re-verify before reuse
These rules protect more than inbox placement. They protect scale. A system that relies on someone remembering which CSV to clean will fail as volume increases.
Good monitoring protects revenue, not just reputation
Deliverability work gets framed as defensive. That is the wrong view.
The same discipline that protects inbox placement also protects pipeline quality. Clean suppression, provider-level monitoring, and fast troubleshooting keep strong domains active and bad segments contained. That lets you scale the parts of outbound that work instead of burning the full setup because one campaign went sideways.
This is the difference between reactive sending and resilient infrastructure. Reactive teams keep mailing until a domain is cooked. Operators with a real system see the warning signs, cut the failing segment, preserve the healthy mailboxes, and keep the machine running.
How do you recover a burned cold email domain?
Stop sending from it first, then decide whether it is worth saving. Most guides cover prevention and stop there, and there is no reliable published data on how long a badly damaged domain takes to recover, so the decision comes down to how the damage happened.
The emergency sequence
- Stop all outbound from the domain. No test sends. Every message you send while you look for the cause adds to the record the providers are judging.
- Find the cause. A bad list, a volume jump, broken authentication, weak targeting, spammy copy, or a mix.
- Isolate the damage. Work out whether one mailbox is contaminated or the whole domain is.
- Fix the root cause before touching volume. If the cause is still there, recovery will not happen.
- Re-warm slowly, only if the damage looks reversible, with clean data and low volume while you watch placement.
When recovery is worth trying
Try to save the domain when it still has some trust, the cause is obvious, and the damage came from one short mistake: a bad list that slipped through, a ramp that went too fast, authentication that was wrong and has been corrected, or one campaign that went badly on an otherwise clean domain. Keep volume low, keep the list strict, and check placement every day until it holds.
When to retire the domain
Retire it when the history is long: repeated spam placement, a high complaint record, or a domain inherited with years of aggressive sending behind it. A clean replacement with proper setup and a 4-week warm-up is usually faster than nursing an old one back. Reachly retires any domain that lands on an MXToolbox blacklist instead of repairing it, and keeps reserve capacity so the campaign keeps running while the replacement warms. The warm-up schedule is in how long to warm up a cold email domain.
What should you check before a cold email campaign launches?
Before any new domain or sequence goes live, run through this list. If anything is unchecked, do not send.
- Sending domain is brand-adjacent, not the primary
- SPF, DKIM, DMARC configured and tested with a DNS checker
- DMARC reporting set up and feeding into a mailbox you actually read
- Sending domains warmed for 4 weeks before campaign traffic
- Daily volume per mailbox kept low, with spare mailboxes in reserve
- List verified within 48 hours of send
- Role accounts, catch-alls, and previous bounces suppressed
- Segmentation locked: one persona, one trigger, one offer per cohort
- Daily monitoring dashboard configured for bounce, complaint, provider-level placement, positive reply
- Backup mailbox capacity ready in case one slips
This list is boring. Operators who run it ship campaigns that do not collapse. Operators who skip it usually call us 90 days later when the domain is cooked.
How do the pieces fit into one outbound system?
Cold email deliverability is not one trick.
It is a four-part system: protected domains with proper authentication, disciplined warm-up and controlled scaling, strict list hygiene, and daily monitoring that catches trouble before a domain gets burned.
That is the difference between reactive outbound and durable outbound. Reactive teams wait for open rates to collapse, then scramble. Serious teams build infrastructure that makes collapse less likely in the first place.
This work is operational. It needs process, tool discipline, and someone paying attention every day. For teams that do not want to own all of that internally, Reachly runs done-for-you multichannel outbound across email, LinkedIn, and phone, uses dedicated sending domains and verified data workflows, and manages sequencing, reply handling, and appointment setting so internal teams can focus on sales conversations instead of inbox maintenance.
If you want predictable outbound, treat deliverability like infrastructure. Because that is exactly what it is.
Book a working call if you want this built and run for you instead of managed piecemeal across inboxes, domains, data vendors, and sequencing tools.
Reachly runs this audit as a standalone engagement through its email deliverability services.
Hand your sending infrastructure to Reachly
Reachly sets up and warms the domains, verifies every list, watches placement every week and rotates tired mailboxes out, as part of done-for-you outbound across cold email, LinkedIn and cold calling.
Book a call with Reachly →
FAQ
What is cold email deliverability and why does it matter in 2026?
It is how reliably your cold emails reach the inbox rather than spam or promotions. It matters more now because Gmail's sender rules, in force since February 2024, and Outlook's, enforced since May 2025, made authentication and complaint rates hard requirements rather than good practice, so a setup that got by two years ago now gets filtered.
How many cold emails can I send per mailbox per day?
Fewer than most tools allow. Reachly's 30,000-email month for one client runs across 180 mailboxes, which averages about 8 cold sends per mailbox per working day. When you need more volume, add mailboxes and domains instead of pushing each inbox harder.
How long should I warm up a new sending domain?
Four weeks before campaign traffic touches it, which is the schedule Reachly runs on every new domain. Keep low-level warm-up running after launch too, so the mailbox keeps a normal-looking pattern while campaigns are live.
Should I send cold emails from my main company domain?
No. Use brand-adjacent sending domains. If one gets flagged, your support, hiring, finance and founder email are untouched. Sending cold outreach from the primary domain puts all of that at risk for one bad campaign.
Do I need SPF, DKIM, and DMARC for cold email?
Yes, all three, before launch. Gmail requires SPF or DKIM from every sender and all three from anyone sending over 5,000 messages a day to Gmail, and Outlook enforces the same for high-volume domains. A missing record shows up in your metrics looking like a warm-up problem.
Why are my cold emails going to spam even with good copy?
Because providers score the sender before the message. Incomplete authentication, mailboxes that were never warmed, or a stale list will send excellent copy to spam. Fix the setup and the data, and the same copy usually starts landing again.
Recommended service
Cold Email Agency
Reachly runs cold email end to end: dedicated infrastructure, signal-led copy, and replies handled until a meeting is booked.
