Key Findings
Mailbox providers grade you on technical setup, sending behavior, bounce rate, and reply quality. Break those rules and Google or Microsoft will bury you without a warning. Fix infrastructure before touching copy.
Without all three configured correctly on a separate sending domain, you start every campaign with a trust deficit. Always run cold outbound from a dedicated domain, never from your main company domain.
Elite senders cap volume at around 100 cold emails per day per address and run roughly 15% warming volume alongside cold sending. Rushing the ramp is one of the fastest ways to wreck a fresh domain.
Reachly holds bounce rate under 3% and deliverability above 97% through multi-pass list verification. Average campaigns often run at 7-8% bounce, which signals weak data hygiene that no copy rewrite will fix.
If the cause is a one-off mistake, slow re-entry with clean data can work. If the domain has a long pattern of abuse, replacing it beats waiting weeks for trust that may never return. Stop sending while you diagnose.
You launch a new sequence on Monday. The targeting is tight, the offer is relevant, and by Thursday the campaign looks dead. Open tracking is noisy, replies are thin, and the team starts blaming copy.
Cold email deliverability is usually the main problem.
If your emails are landing in spam, promotions, or nowhere visible, the rest of the outbound stack does not matter. Personalization does not matter. Your sequencing tool does not matter. A better CTA does not matter. You cannot book meetings from messages people never see.
The operators who win treat deliverability like infrastructure, not superstition. Clean sending setups routinely produce inbox placement in the high nineties, while neglected setups can collapse fast and stay there until you fix the root cause. Reply rates are also tighter than they used to be, with fewer mistakes tolerated at every stage of the funnel (source).
That is why this guide covers the full lifecycle, not just setup. It covers how to prevent problems, how to monitor them early, and how to recover when a domain is already damaged. That last part gets skipped in a lot of deliverability advice because recovery is messy, slow, and operationally inconvenient. It is also the part founders need when an outbound engine suddenly stops producing.
The job is simple to describe and hard to execute. Build trust with mailbox providers, protect that trust while you scale, and know how to repair it when you burn it.
Why Your Emails Are Landing in Spam (And How to Fix It)
The first bad assumption is that spam placement means your offer is weak. Usually it means the mailbox providers do not trust you.
That trust gets built from signals. Technical setup. Sending behavior. Bounce rate. Reply quality. Whether people ignore you. Whether your pattern looks like a human operator or a machine firing junk into the market. If you break those rules, Google and Microsoft do not warn you. They just bury you.
What inbox providers actually care about
Cold email is harder than newsletters because the recipient did not ask for it. That means small mistakes hurt more. Mailbox providers already expect lower engagement, so they watch your setup and behavior more closely.
Three things matter most. DNS setup, where SPF, DKIM, and DMARC need to be correct. Warm-up discipline, where new domains get time before real volume. Stable sending, where sudden spikes make you look fake.
If you want a plain-English primer on the mechanics behind inbox placement, this is a useful reference because it breaks down the infrastructure side without the usual fluff.
The first fixes to make
Start with your infrastructure, not your copy. Copy tweaks will not save a broken domain. Then look at your subject lines. Not because subject lines control deliverability on their own, but because weak ones can crush engagement and make a borderline setup worse. If you need a reset there, Reachly's guide on cold email subject lines is worth reading after you have fixed the technical basics.
The short version is simple. Deliverability is not mysterious. It is a scoring system, and you are either feeding it good signals or bad ones.
The Essential Technical Foundation
A lot of deliverability problems start before the first email goes out. The domain is fine on paper, but the records are wrong, the tracking setup is sloppy, or the sending domain was picked without checking its history. Then founders blame copy, offers, or list quality. Fix the setup first.
SPF, DKIM, and DMARC in plain English
These records decide whether inbox providers treat your mail as legitimate or suspicious. SPF lists which servers can send on behalf of your domain. DKIM signs the message so providers can verify it was not altered. DMARC tells providers what to do when SPF or DKIM fails, and it gives you reporting so you can spot problems.
If one is missing, misaligned, or broken, you start with a trust deficit. That matters even more in outbound, where engagement is already weaker than opt-in email and reputation is harder to rebuild once it slips.
Essential Email Setup Checklist
The setup checklist that actually matters
Before launch, verify the infrastructure from the live domain and live mailbox, not from a setup doc.
Use a separate sending domain. Keep cold outbound off your main company domain. If the sending domain gets flagged, you can replace it. Your support, hiring, and investor email stay clean. Check SPF, DKIM, and DMARC in a real tester like Google Admin Toolbox, MXToolbox, or Mail-Tester. Confirm the records resolve, align with the visible From domain, and pass from the actual mailbox you plan to use. Set a basic DMARC policy from day one. Start with p=none for visibility if you need reporting first, then tighten later once you know all legitimate mail sources are covered. Match your mailbox provider and sequencer settings, so if you send through Google Workspace or Microsoft 365, make sure Smartlead or your sending platform is authenticated correctly and not creating avoidable misalignment. Test placement before real volume by sending seed emails to Gmail, Outlook, and Apple Mail inboxes you control. Choose sane domain variants, since something like getbrand.co or brandmail.co is defensible while obvious throwaway patterns make both prospects and filters suspicious.
A lot of outbound teams skip the domain history check. Bad call. Expired domains, spammy ownership history, and weird registration footprints can make recovery harder later (source).
What to avoid
Do not load five or ten mailboxes onto a fresh domain on day one. Do not mix cold outbound with normal business email. Do not let an SDR connect a mailbox to a sequencer before someone checks authentication, custom tracking, unsubscribe handling, and test sends.
Those shortcuts create the exact kind of damage that is annoying to diagnose and expensive to recover from. Strong copy cannot rescue a domain that fails authentication.
The trade-off founders usually resist
Separate domains create extra work. You need to buy domains, provision mailboxes, set DNS, verify records, and monitor reputation. That overhead is real. It is still cheaper than repairing a damaged primary domain.
We have seen teams save an hour on setup, then lose weeks trying to recover inbox placement after running outbound through their main company domain. If this guide has one point that carries all the way through monitoring and recovery, it is this. Treat deliverability like infrastructure from the start, or you will end up doing cleanup under pressure later.
The Domain Warm-Up Playbook
New domains have no trust history. You cannot throw volume at them and hope for the best. That is the cold email version of hiring a rep on Monday and making them close enterprise deals by Tuesday. The mailbox exists, but the reputation does not.
The operating limits that matter
Elite B2B senders keep 95%+ inbox placement by using 1 to 5 dedicated sending addresses per domain, capping volume at around 100 cold emails per day per address, and running roughly 15% warming volume alongside cold sending. Autopilot warm-up tools also send around 50 engaging emails per day per address before cold sending begins (source).
That does not mean you should instantly open five mailboxes and hit the cap. It means the infrastructure has a ceiling. Rushing to that ceiling is how teams wreck it.
A warm-up schedule that will not get you punished
Use a tool like Smartlead, Instantly, or a similar warm-up system connected to real inboxes. The specific tool matters less than the discipline.
A four-week warm-up plan for a fresh sending domain. Each phase has a clear job and a clear failure mode to avoid.
| Week | Phase | What to do | What not to do |
|---|---|---|---|
| 1 | Warm only | Run warm-up traffic only. Watch inbox placement and spam folder behavior on seed tests. | Do not send any real campaign traffic yet, even to friendly prospects. |
| 2 | Light send | Add a small amount of real outbound while warm-up continues. Keep daily volume flat. | Do not ramp because one good day made the dashboard look healthy. |
| 3 | Controlled ramp | Increase carefully if placement stays clean and replies look normal across providers. | Do not spike volume after early wins. One bad day will undo three good ones. |
| 4 | Hold steady | Keep volume stable. Consistency builds reputation faster than squeezing extra sends. | Do not rotate in random lists just to fill capacity. |
How Reachly runs this at scale: Domains are provisioned and managed through ZapMail, sequences run inside Smartlead with mailbox rotation and reply tracking, and lists are built and segmented in Clay using buying signals like hiring, funding, and tech stack changes. LinkedIn touches run in parallel through HeyReach on the same account list so channels do not collide.
A common mistake is treating warm-up like a countdown timer. It is not. It is a reputation-building process. If tests show weak placement, stay slower for longer.
How many mailboxes to use
For most early outbound programs, fewer mailboxes are better. More mailboxes create more moving parts, more monitoring, and more ways to mess up reputation.
Use a small set of dedicated inboxes. Keep daily patterns stable. Do not send all volume in one burst at the top of the hour. Randomized pacing helps you look less robotic, but it does not save bad fundamentals.
Warm-up works when the rest of your setup is clean. It does not rescue a junk list or bad domain hygiene.
Tools people use badly
Smartlead is useful for mailbox rotation and warm-up. HeyReach is useful when you are coordinating email with LinkedIn workflows. Instantly is common for quick deployment.
But tools do not create trust by themselves. If your domains are sloppy, your lists are stale, or your send volume jumps around like crazy, the tool just helps you fail faster.
List Hygiene and Content That Gets Replies
Bad data wrecks sender reputation faster than mediocre copy. Founders usually assume the opposite because copy is more visible. It is not the bigger problem. Sending to invalid contacts is.
Bad lists poison everything
Bounce rate should sit under 2%, while average cold campaigns often run at 7% to 8%. Teams using multi-provider verification and fixing list quality have moved inbox rates from 65% to 92% in roughly one month (source).
Reachly's own benchmark is similar. We hold bounce rate under 3% and deliverability above 97% across client campaigns. The rule is the same either way. If your data is weak, your deliverability falls apart even if the rest of the system is decent.
Use ZeroBounce, MillionVerifier, or a multi-pass verification workflow before a lead enters a sequence. One verifier is better than none. Multiple checks are better than blind trust. In our stack, lead lists run through a verification waterfall inside Clay before anything is loaded into a sequence.
What to clean before a lead goes live
A clean list usually comes from boring process, not clever tech. Verify every address rather than assuming a scraped email is safe because one tool found it. Remove stale records, since old contacts create bounces and low engagement. Segment tightly, because a narrow list with relevant messaging beats a broad list with weak fit. Check role accounts carefully, since generic inboxes often behave differently from named contacts. Suppress problem contacts, so if someone bounced, complained, or is clearly not a fit, you stop retrying.
This is also where message quality starts to matter. If people ignore you, that hurts. If they reply, that helps.
Write for replies, not applause
Short emails tend to be easier to read and easier to answer. Reachly's own rule for 2026 is 70 to 80 words per cold email with a single low-friction CTA. Long product tours get skimmed or deleted. A solid cold email usually does three things. It shows you know who the person is. It mentions a problem they likely care about. It makes the reply easy.
The reply rate benchmark to aim for is 10 to 20% positive replies for a well-run B2B campaign, with 35 to 40% being exceptional. If you are sitting under 5%, the issue is usually one of four things: deliverability, subject line, message length, or a weak offer.
If you need examples of that style, Reachly's B2B cold email templates are a better starting point than the usual "hope you are well" filler. The sequence those emails run inside, including how many follow ups to send and when to stop, is in our cold email follow up playbook.
How to Monitor and Diagnose Deliverability Issues
If you are not checking placement, you are guessing. Guesses are expensive. Most operators look at reply rate and call it a day. That is too late. By the time replies collapse, the damage is usually already visible in placement, opens, and provider-specific behavior.
What to watch every week
You need a short list of checks, not a giant dashboard nobody reads. Inbox placement tests, run through seed testing tools before major launches and after any major volume or list change. Bounce behavior, where any rise in bounce rate stops the copy debate immediately. Provider splits, comparing Gmail and Outlook behavior, since if one provider tanks that points to a filtering issue there. Reply quality, because short meaningful replies matter more than vanity opens. Sending consistency, since sudden volume changes often show up in performance before anyone notices the root cause.
A lot of teams use Smartlead or similar sequencers for campaign management, but the primary value is in watching mailbox-level performance instead of only campaign-level totals.
A simple diagnosis table
| Symptom | Likely issue | First move |
|---|---|---|
| High bounces | Bad data or poor verification before sending | Pause the list and re-verify contacts through a multi-step waterfall. |
| Low opens, one provider | Reputation issue specific to Gmail, Outlook, or another provider | Check recent volume, run placement tests, audit domain health. |
| Low opens, all providers | Infrastructure or domain-wide problem | Audit setup end to end. Pause campaigns until the cause is clear. |
| Replies weak on broad lists | Relevance and segmentation problem, not a sending problem | Tighten segmentation. Rewrite messaging to match the segment. |
| Spam complaints rising | Wrong audience or weak message fit triggering negative signals | Cut bad segments fast. Suppress complainers permanently. |
Why intent matters more than most deliverability guides admit
Most deliverability advice talks about list quality but stops short of intent. That is a miss. Outreach to high-intent accounts tends to produce higher reply rates, and replies are a core positive signal for inbox providers (source).
In practice, that means you should track segments separately. Leads with hiring spikes, funding activity, or obvious change events should not be mixed into the same performance bucket as generic scraped lists. This is exactly how Reachly runs campaigns. Lists are segmented by signal in Clay, pushed into Smartlead on dedicated sending domains, and coordinated with HeyReach for LinkedIn so each segment has a clean performance read. When the funding-triggered segment outperforms the hiring-triggered segment by 4x, we know it instantly. Most teams do not, because their performance data is averaged across mismatched segments.
What not to do during diagnosis
Do not keep sending while watching what happens. That usually makes things worse. And do not change five variables at once. If you rewrite copy, swap domains, load a new list, and alter volume on the same day, you will not know what fixed the issue.
The Recovery Protocol When Your Domain Is Burned
This is the part most guides skip. They tell you how to avoid trouble. Fine. Useful. But if your domain is already cooked and your opens are wrecked, that advice arrives too late.
Most cold email content explains prevention but offers very little on recovery timelines or whether a badly damaged domain can be fully saved. The 65% to 92% inbox improvement example often quoted in deliverability content is about fixing basics, not recovering from a serious reputation hit (source).
The emergency sequence
When a domain is burned, do this in order. Stop all outbound from that domain. Do not test a few more sends. Stop. Check what caused it: bad list, aggressive volume, broken setup, weak targeting, spammy copy, or some ugly mix of all five. Separate infrastructure to figure out whether one mailbox is contaminated or whether the whole domain is compromised. Fix the root issue before touching volume, because if the cause stays in place, recovery will not happen. Re-warm slowly only if the damage looks reversible, with controlled traffic, clean data, and zero impatience.
When recovery is worth trying
Recovery makes sense when the domain still has some trust left, the root cause is obvious, and the problem came from a short-term mistake rather than a long history of abuse. Examples include a bad list that slipped through and caused bounce issues, volume that jumped too quickly during early ramp, authentication that was misconfigured and later corrected, or one campaign that went off the rails while the domain stayed mostly clean.
In those cases, slow re-entry can work. Keep volume low. Keep list quality strict. Watch placement constantly.
When to retire the domain
Sometimes the honest answer is to bin it. If the domain has a long pattern of abuse, repeated spam placement, ugly complaint history, or you inherited an account with years of outbound debt, trying to save it can waste more time than starting over. There is no clear published data on exact recovery windows, which is why operators need judgment here.
A clean replacement domain with proper setup and disciplined warm-up is often the faster move. It feels annoying because you lose continuity. It is still the right move when the old infrastructure keeps poisoning new campaigns.
If you are weighing whether to rebuild the stack internally or hand it to a specialist, Reachly's overview of what a cold email agency actually does can help frame what managed infrastructure and monitoring should entail.
This is what Reachly handles end to end. Dedicated sending domains and mailboxes provisioned through ZapMail and kept off the client's primary domain. SPF, DKIM, and DMARC verified before launch. Lists built and segmented in Clay using buying signals like hiring, funding, and tech stack changes. Sequences run through Smartlead with mailbox rotation and weekly placement monitoring. LinkedIn coordinated through HeyReach on the same account list, with cold calling layered on for high-value accounts. When something breaks, infrastructure gets isolated and rebuilt without your sales team losing a day to DNS records.
Get more meetings with the people who matter, 100% done for you.
We do not spray and pray. We use real buying signals to reach the right people at the right time, then run coordinated outreach across email, LinkedIn, and phone with messaging that earns replies.
Get Started
Frequently Asked Questions
SPF authorizes which servers are allowed to send email on behalf of your domain. DKIM cryptographically signs each message so receiving servers can confirm it was not altered in transit. DMARC ties the two together by telling inbox providers what to do when SPF or DKIM fails, and it sends you reports on what is passing or failing. All three need to be configured correctly before you start sending cold outbound.
No. Always use a separate sending domain for cold outbound. If something goes wrong, whether it is a bad list, a bounce spike, or a reputation hit, you can replace the domain without affecting your support, hiring, billing, and investor email. Mixing cold outbound with your main company domain is one of the fastest ways to damage critical email infrastructure.
Plan for two to four weeks. Week one is warm-up traffic only with no real sends. Week two adds a small amount of real outbound. Weeks three and four ramp carefully if placement and reply quality look healthy. The exact timeline depends on the warm-up tool, the daily volume target, and how clean the early sending behavior looks. Treat it as a reputation-building process, not a countdown timer.
Aim for a bounce rate under 3% and inbox placement above 97%. Reachly holds these benchmarks across client campaigns through multi-pass list verification and disciplined sending behavior. Average cold campaigns often run at 7% to 8% bounce rate, which signals weak data hygiene. On reply rate, 10 to 20% positive replies is normal for a well-run B2B campaign with proper infrastructure and signal-based targeting, while 35 to 40% is exceptional.
It depends on how badly it was damaged and what caused it. If the issue was a one-time mistake like a bad list or rushed warm-up, slow re-entry with strict list quality and controlled volume can work. If the domain has a long history of spam placement, complaints, or repeated abuse, retiring it is usually faster than recovery. There is no clean published data on exact recovery windows, so judgment matters. When in doubt, replace the domain rather than waste weeks waiting for trust that may never come back.
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.
