Deliverability and infrastructure

Deliverability is the share of your emails that reach the inbox instead of spam, a promotions tab or nowhere at all. For cold outreach it decides everything downstream, because a message nobody sees cannot get a reply. Nothing warns you when it slips. The dashboard shows every email as sent, the replies dry up, and the team starts rewriting subject lines while the real problem is a flagged domain. That is why the sending side of our cold email service gets built before anyone writes a word of copy.

Open rate tries to measure whether someone looked at an email; deliverability is whether it reached the inbox at all. Opens are counted through an invisible tracking image. Since Apple Mail Privacy Protection arrived in 2021, many inboxes load that image automatically, while other mail apps block images and hide real opens. So open rate is a rough hint at best. We judge cold email results on replies, conversations started and calls booked, because a mail app cannot inflate those.

In failed campaigns, the causes show up in a predictable order. Sending from the main company domain. Missing SPF, DKIM or DMARC records. No warm-up, or three days of it because somebody was impatient. One mailbox doing the work of twenty. Unverified lists bouncing hard enough to undo the warm-up. The same block of text sent ten thousand times with no variation. Spammy wording matters too, but it rarely comes first. Every item on that list is a setup mistake, and in the campaigns we run each one is ruled out before launch.

Inbox placement is mostly decided before the first send. Smirnov Consulting Group registers new domains so your main one is never exposed, authenticates each of them, warms every mailbox before it touches a prospect, and keeps volume per mailbox low enough that each account still behaves like a person writing email. Every address is verified first, and messages go out with spintax variants so sends do not repeat word for word. After launch, running cold email for a client mostly means keeping all of that true as volume grows.

Always. Cold campaigns never go out from the domain your invoices, contracts and support replies come from, and that includes subdomains like outreach.yourcompany.com, which still sit on your root domain. For cold email outreach we register separate domains instead, usually close variants of your brand, pointed somewhere real rather than at a blank page. A serious campaign typically runs on 10 to 20 of them, with mailboxes spread across each.

Not when it is set up properly, because your main domain never sends a single cold email. If it did, the risk would have no upper limit: once the main domain is flagged, your everyday business email starts landing in spam next to the outreach, and repairing that is slow. With all cold volume on other domains, the worst case is losing one sending domain. It gets replaced, and your company inbox never notices. That separation is the first thing Smirnov Consulting Group puts in place on any cold email program.

Yes, on every sending domain, before a single message goes out. SPF lists the servers allowed to send mail for the domain. DKIM adds a cryptographic signature showing the message really came from it and was not altered. DMARC tells receiving servers what to do when those checks fail and sends back reports. Without them you are a stranger claiming to be a company, and since 2024 Gmail and Yahoo require authentication from bulk senders anyway. It is about an hour of DNS work, and on a done-for-you cold email build our team handles it.

Warming means a new mailbox sends small volumes to real inboxes and gets replies, building up slowly until it looks like an account a person actually uses. Ten days is the minimum we accept. On most cold email builds the accounts get 10 to 14 days before they touch a single prospect, and a brand new domain runs on the same clock, since its mailboxes are what get warmed. Founders hate this step, because they are paying while nothing visible happens. Skip it, though, and the first campaign lands in a folder nobody opens.

Plan on roughly 25 emails per mailbox per day. Twenty-five, not two hundred. That figure comes from a real campaign for a UAE tech startup: about 1,250 emails a day sent from 50 email accounts, which booked 684 demos in 4 months. Spread the sends evenly across the working days rather than in bursts. When a campaign needs more daily volume, the answer is more mailboxes, never pushing each one harder.

Start from the goal and work backwards to the mailbox count. For 30 to 35 sales calls a month from outbound, you need infrastructure that can reach around 1,000 people a day. That usually means 20 to 40 mailboxes spread across 10 to 20 domains, more for a large company with a big market, and every cold email build we run is sized this way. Two mailboxes and a thousand sends a day is a bonfire.

Inbox rotation means spreading one campaign's sends across many mailboxes and domains, so each account carries a small share instead of one address doing all the work. Providers judge each mailbox and domain on its own behavior, and one overloaded account gets flagged quickly. A financial advisory client in Australia ran on 20 email accounts: every rep's cold volume rotated across three mailboxes, with a fourth kept for follow-ups.

Every address is verified before it enters a campaign, and only verified ones get sent to. Finding comes first: Smirnov Consulting Group pulls contacts from more than 16 B2B data sources. A found email is still a guess until it is checked, so each one goes through NeverBounce, which beat the other verifiers we tested, and sometimes DeBounce. On every campaign we manage, anything that fails is dropped rather than sent and hoped for.

Under 0.5% is the bar we hold every campaign to. Above that, bounces start undoing weeks of careful warm-up, since they are one of the fastest ways a domain gets flagged. The number stays low through verification before sending, much more than through fixes afterwards. If bounces climb mid-campaign, the sensible response is to pause the affected sequence, find the batch that caused it and re-verify it before anything else goes out.

Not free consumer accounts. A personal @gmail.com or @outlook.com address has tight sending limits, looks amateur to a B2B buyer and ties the outreach to one person, not your company. Cold email at scale runs on business mailboxes on domains you own, authenticated and warmed before use, which is how we set up cold email for clients. The familiar Gmail or Outlook interface is fine; the account type and the domain behind it are what count.

Business mailboxes on Google Workspace or Microsoft 365, on dedicated sending domains, with a sequencing tool on top. That is the standard setup for cold outreach. Newsletter platforms are the wrong fit. They are built for opt-in subscriber lists, and most of their terms forbid emailing people who never signed up. Between the two big mailbox providers, the choice matters less than what sits underneath: authentication, warm-up, modest volume per mailbox and verified data decide placement.

Campaigns send through dedicated sending software we run for each client, from business mailboxes on domains our cold email agency sets up for each client, with spintax variants in every sequence. Of the whole chain, the sending tool is the least important link. Teams that blame the software usually skipped the domains, the warm-up or the data. Move a weak setup to a better platform and you get the same spam folder with a different logo.

Yes. Every client and every campaign sends from dedicated IP addresses, with separate IPs even across different domains, not just across clients. A single client can run on 10 to 20 different sending IPs at once to keep deliverability safe as volume grows. Mailbox and domain reputation still carry most of the weight day to day, but isolating IPs stops one domain's problems from bleeding into another's. In a typical cold email setup, that isolation sits alongside mailbox count as one of the controls we manage.

The Smirnov Consulting Group team sets up the whole infrastructure, from registering the sending domains to verifying the list before anything goes out. Most companies would rather never think about DNS records, which is a perfectly reasonable preference. What we need from you is someone ready to take the sales calls once meetings start landing. Building it in-house is possible too, since nothing on the checklist is secret. It takes patience more than skill, mostly during warm-up.

We usually stop a cold email sequence at four steps, and deliverability is part of the reason. Every follow-up is another send from the same mailbox, so sequence length and deliverability share one budget. Three emails to each of 5,000 people is 15,000 sends, and every one of them counts against a mailbox's daily limit of about 25. Later emails with nothing new to say are also the ones people mark as spam, and a complaint does more harm than silence. So every step in the cold email sequences Smirnov Consulting Group builds has its own job.

By keeping volume per mailbox flat and scaling sideways. When a campaign works, we add mailboxes and domains instead of asking existing ones to send more, and new contacts are verified batch by batch as they go in. So a mailbox sending around 25 emails a day in month one is still sending around 25 in month six, and the extra reach comes from the accounts added beside it. That discipline is what keeps our longer campaigns in the inbox rather than any single trick.

Mostly through the replies, because direct complaint data is thin in cold email: few sending domains send enough for providers to report on them. Every reply is read and qualified by our team, so a wave of annoyed answers or removal requests shows up fast. Google Postmaster Tools shows spam rates for a domain once volume is high enough, and Google's own guidance is to stay under 0.1% and never reach 0.3%. A sudden drop in replies from one mailbox is the other early warning.

Daily during the first weeks after a cold email launch and after any increase in volume, then at least weekly once a campaign is stable. Early checks matter most, because a mailbox that starts sliding keeps sending, and most teams only look once the replies have already stopped. Watch bounce and reply rates per mailbox, blocklist status for each domain, and spam-rate data where a provider publishes it. A drop confined to one mailbox points to that account. A drop everywhere points to the list or the copy.

The campaign keeps running, because properly built cold email never depends on one domain. A blacklisting takes out a slice of capacity while the other domains carry on. Losing a sending domain is annoying; losing your main one is what the whole setup exists to prevent. The usual response is to stop sending from the flagged domain and find the cause first, most often an unverified batch or an overloaded mailbox. Recovery means requesting delisting where the blocklist allows it, resting the domain and warming it back up slowly. If it does not recover, it gets replaced.