Every time a company comes to us with a deliverability problem, we ask the same diagnostic question: "What have you changed about your infrastructure in the last 90 days?" The answer is almost always the same. They haven't changed anything. They've been A/B testing subject lines, experimenting with preview text, shortening their sequences, hiring better copywriters - all the things every cold email guide tells you to do. And they're still landing in spam.
Here's what the SaaS tools and content farms won't tell you: subject lines are a rounding error in deliverability. Gmail and Outlook's filtering algorithms are primarily making decisions about your email before they ever process your content. They're looking at your domain's age and reputation, the pattern of your sending volume, and whether your authentication records tell a consistent, trustworthy story. If any of those three signals are broken, your perfectly-crafted subject line never gets seen. You can write the most compelling opening line in B2B history and it still lands in spam. That's the conversation we need to have.
The Real Reason Your Emails Land in Spam
There are dozens of factors that influence deliverability, but in practice, the campaigns we audit almost always fail for one of three structural reasons. These aren't edge cases - they're the default state of most cold email setups built without infrastructure expertise.
Google and Microsoft both score your sending domain based on its history. A domain registered last month with no prior email activity, no organic traffic, and no inbox engagement is a red flag by definition. Spam operations register domains in bulk and start blasting immediately - so inbox providers have trained their systems to treat fresh domains as suspects. If you bought a new domain and started sending 200 emails per day from it within a week, you've replicated exactly what spammers do. The infrastructure fix isn't clever - it's patience: domains need a minimum 60 - 90 day aging and reputation-building period before they can sustain serious outreach volume. Many teams skip this because no one told them it was required.
Even a seasoned domain with good reputation will get flagged if you violate volume ramp norms. Spam filters watch for sudden volume spikes because they signal acquired or harvested lists. The safe ramp from a new mailbox is roughly: 5 - 10 emails on day one, increase by 10 - 15% per day, never exceeding 50 emails per day per mailbox in the first 30 days. Most teams skip straight to 200 - 500 per day on a new mailbox - and wonder why deliverability collapses. The math is simple: if you need to send 1,000 emails per day, you need a rotating pool of 6 - 8 warmed mailboxes, not one mailbox in overdrive. Volume is an infrastructure problem, not a courage problem.
SPF (Sender Policy Framework), DKIM (DomainKeys Identified Mail), and DMARC (Domain-based Message Authentication) are DNS records that prove you are who you say you are. SPF lists which servers are authorized to send on your domain's behalf. DKIM cryptographically signs every outgoing message so the recipient's server can verify it wasn't altered in transit. DMARC tells receiving servers what to do if a message fails SPF or DKIM checks - reject it, quarantine it, or let it through with reporting. Without all three properly configured, you are functionally anonymous to inbox providers. In 2026, Gmail's bulk sender policies require valid DKIM and DMARC before any large-volume outreach is even considered for inbox placement. Most teams have SPF set up - it's the other two where they fall down, especially DMARC policy (which should be set to at minimum p=quarantine to signal you're actively managing your domain's authentication).
One important distinction: these are structural failures, not content failures. They exist independently of what your email says. Fixing your subject line while your SPF record is misconfigured is like replacing the carpet on a sinking ship.
The 5-Layer Infrastructure Stack That Actually Fixes Deliverability
When we onboard a new cold email client at Deep-Y, we don't touch a single sequence until the infrastructure is locked. There are five layers - each one is a prerequisite for the next. Here's the full stack, in order:
-
01
Dedicated Sending Domains
Never send cold outreach from your primary company domain. A deliverability problem on your outreach domain should not put your main domain's reputation at risk - they need to be separated. Set up 2 - 4 dedicated sending domains that are variations of your main brand (e.g. trydeep-y.com, deep-y.io, getdeep-y.com). Configure each with its own DNS records, independent of the others. This also allows you to A/B test domain performance and rotate out burned domains without interrupting active campaigns.
-
02
Mailbox Rotation Pool
Assign 2 - 3 mailboxes per sending domain (e.g. maor@trydeep-y.com, hello@trydeep-y.com, team@trydeep-y.com). Your sending tool distributes outbound volume across all active mailboxes, so no single mailbox is hitting volume thresholds. This gives you 6 - 9 active mailboxes across 3 domains, safely capable of 300 - 450 emails per day total once warmed - without any single mailbox raising red flags. Tools like Instantly, Smartlead, or Reply.io all support mailbox rotation natively.
-
03
Structured Warm-Up Sequence
Every new mailbox runs through a 4 - 6 week automated warm-up before it touches real prospects. Warm-up services (Warmup Inbox, Mailreach, or built-in warm-up within Instantly/Smartlead) send small volumes of real back-and-forth email activity between a network of inboxes - creating the engagement signals that tell inbox providers this is a legitimate, active account. Set warm-up volume to start at 10/day and auto-increment by 10 - 15% weekly. Critically: warm-up should continue running in the background even after you start live campaigns - typically 30 - 40% of your total sending volume should always be warm-up traffic to maintain reputation signals.
-
04
Full Authentication Setup (SPF + DKIM + DMARC)
For each sending domain, configure SPF to authorize only your sending server (e.g. Google Workspace, SendGrid, or your ESP). Set up DKIM with a 2048-bit key - not the default 1024-bit, which is increasingly flagged by modern spam filters. Configure DMARC at a minimum of
p=quarantine, and point theruatag to a reporting inbox so you can monitor authentication failures in real time. Also add a custom tracking domain for your click tracking (e.g. track.trydeep-y.com) to avoid sharing a tracking domain with other senders - shared tracking domains are a common, overlooked reputation contaminant. This entire DNS setup takes 30 - 60 minutes per domain when you know what you're doing; it takes most teams days of back-and-forth with their hosting provider because they've never done it before. -
05
Ongoing Reputation Monitoring
Infrastructure isn't a one-time setup - it requires active monitoring. Check your domains weekly in Google Postmaster Tools (free), which grades your domain reputation as high, medium, low, or bad. Run your sending IPs through MXToolbox's blacklist checker before each new campaign launch. Monitor your bounce rate (hard bounces should never exceed 3% of sends) and spam complaint rate (Gmail flags anything above 0.1%). When any signal degrades, you need a protocol: pause, diagnose, rotate to a backup domain, never blast through it. The teams that maintain 85 - 90% open rates long-term are the ones who treat infrastructure monitoring as a weekly operational habit, not a once-and-done DNS task.
This stack requires legitimate technical depth to implement correctly. It's not prohibitively complex - but it's not something you set up in an afternoon with a YouTube tutorial either. And the cost of getting it wrong isn't just low open rates; it's domain blacklisting that can take months to recover from.
The Numbers When You Get It Right
We've built this stack for B2B companies across HVAC, SaaS, solar, data intelligence, and immigration services. In every case, the deliverability transformation is the same: moving from industry-average 20 - 25% open rates to well above 80%. These aren't projection numbers - they're actuals from live campaigns.
AirCentral is a commercial HVAC company that had essentially no existing cold email infrastructure. When they came to us, they were sending manually from their primary domain, with no DKIM, no custom tracking domain, and a DMARC record set to p=none. Open rates were hovering around 22%. We rebuilt their entire sending infrastructure from scratch - four dedicated domains, eight mailboxes, full authentication, custom tracking, and a 6-week warm-up before a single prospect was touched. The result was 89% open rate across their outbound campaign, which generated 340 booked commercial meetings and $540K in signed contracts in 90 days. The open rate number didn't come from better subject lines. It came from the fact that their emails were now reliably landing in inboxes instead of spam folders.
Solar Direct was a similar situation - a B2B solar installation company sending from their main domain with a shared ESP tracking link. After we rebuilt their infrastructure with dedicated sending domains, a proper warm-up, and isolated tracking, they hit 85% open rate within the first 30 days of live outreach. Seemore Data, a data intelligence firm, reached 90% open rates by the second campaign cycle. The pattern is consistent: when the infrastructure is correct, open rates above 80% are the norm, not the ceiling.
For context on what higher open rates mean downstream: AirCentral's 89% open rate meant 89 out of every 100 target decision-makers actually saw their message. At 20% open rate, you're effectively invisible to 80% of your list. If you want to understand how we pair this deliverability foundation with AI-powered lead generation and personalization, the full system is explained on our services page.
The DIY vs. Done-For-You Question
You can build this yourself. Every tool mentioned above is publicly available, Google Postmaster is free, and the DNS records are just DNS records. The difference between DIY and done-for-you isn't access to secret knowledge - it's whether your team will spend 40 - 60 hours in the first 90 days doing this correctly, and then maintain the monitoring habit permanently after that. Here's what realistic DIY execution actually costs:
| Task | DIY Reality | Done-For-You |
|---|---|---|
| Domain + mailbox setup | 8 - 12 hrs (first time) | 2 - 3 days, handled |
| SPF + DKIM + DMARC per domain | 2 - 4 hrs each, often requires DNS back-and-forth | Done in setup sprint |
| Warm-up (active management) | 6 weeks, requires daily checks | Automated + monitored |
| Blacklist monitoring | Manual weekly audit, easy to forget | Automated alerts |
| Domain rotation (when one burns) | Reactive, often costs 2 - 4 weeks of downtime | Pre-built rotation pool |
| Ongoing reputation maintenance | Easy to deprioritize under workload | Weekly operational habit |
| Total time investment (first 90 days) | 40 - 60 hours + opportunity cost | Zero from your team |
If you're a founder or sales ops lead with real bandwidth, a methodical approach, and the patience to do it right the first time - DIY is a reasonable path. Budget 40 - 60 hours for the first 90 days, read Google's bulk sender guidelines in full, and commit to the monitoring habit permanently.
Most B2B companies, though, don't have that time. They have a pipeline target and a quarter that won't wait six weeks for their mailboxes to finish warming. The cost of the wrong setup - blacklisted domains, months of reputation recovery, a sales team running on broken infrastructure - typically exceeds the cost of having it built properly by an order of magnitude. The calculation isn't "can we do this ourselves?" It's "what is 60 hours of the right person's time actually worth, and what happens if we get it wrong?"
If you want to go deeper on how infrastructure fits into a full outbound system - lead sourcing, personalization, sequence strategy - our AI lead generation service page walks through the full picture. Or if you're ready to talk specifics, book a free strategy call and we'll audit your current setup and tell you exactly what's broken.
If you're considering whether to replace your SDR team entirely or augment with AI, the deliverability layer is only part of the picture - see our AI SDR vs Human SDR comparison for the full cost and pipeline breakdown. And if you want a concrete example of what properly-built deliverability infrastructure enables at scale, see how we used it to generate 18,000 leads at $0.30 each.
Want this built and running
in two weeks?
That's exactly what we do at Deep-Y. We build the full cold email infrastructure stack - dedicated domains, mailbox rotation, warm-up, authentication, reputation monitoring - then run it for you continuously. No DIY, no guesswork, no six-week ramp before you see results from your own effort.