Most growing businesses pay for email three times — inboxes, campaigns, and verification. The self-hosted stack that replaces all three, and what it costs.
Most growing businesses are paying for email three times over without realising it.
There's the business email — the day-to-day inbox for every team member. There's the campaign platform — the tool that sends newsletters, onboarding sequences, and outbound. And increasingly there's the verifier — the service that cleans lists before sending so deliverability doesn't collapse.
Three vendors. Three invoices. All of them growing as the business grows, and none of them owned by the business.
A self-hosted email infrastructure consolidates all three onto one server the business controls. This is what we deploy and what it replaces.
The team that usually feels this first
It's almost always a 30–80 person company doing real outbound. Marketing is running newsletters to a list of 30,000–100,000. Sales is sending sequences. Sometimes there's a community list on top. Somebody finally added up Google Workspace + Mailchimp + a verifier and the total made everyone in the room go quiet.
That's when the conversation usually starts. Not because anything is broken — because the costs scale with the list, not with the value the business is getting.
Business email — Mailcow
Mailcow is the foundation of every email deployment we do. It handles transactional email — inboxes, SMTP, IMAP, calendar, contacts — across every domain on the account. This is what replaces Google Workspace or Microsoft 365 at the per-seat level. (We covered the full Google Workspace replacement stack in detail here.)
Before a single production email moves over, we configure SPF, DKIM, and DMARC, and verify the records propagate correctly. These three DNS records are the difference between email that lands in inboxes and email that lands in spam. Most self-hosted email failures happen here — the server is running fine, the records aren't quite right, and deliverability is damaged before anyone notices.
We run delivery tests against Gmail, Outlook, and Apple Mail before any cutover. Then both the old and new systems run in parallel for two weeks. From the user's perspective, nothing changes until we've confirmed everything works. rootpath.studio's email moved this way without a single disruption — the case study walks through what that looked like in practice.
Bulk sending — Listmonk with Amazon SES
This is where the cost difference gets loud.
Listmonk handles lists of any size, manages segments and unsubscribes, and exposes a clean interface a marketing team can actually use. It doesn't charge per subscriber, and it doesn't charge per send. It's software running on a server. The cost is the server.
For the actual delivery we route outbound through Amazon SES. At $0.10 per thousand emails, a business sending 500,000 emails a month pays $50 to send them. The same list on Mailchimp Standard at that contact volume runs well over $400 a month — and the gap widens as the list grows.
We handle the full delivery setup: SES verified domains, bounce processing, unsubscribe automation, feedback loop registration with the major inbox providers, and IP warmup so a new sending reputation builds gradually instead of starting cold. Done wrong, warmup damages deliverability for months. Done right, it's invisible.
Email verification — self-hosted Reacher
For businesses doing outreach at any volume, verification is the line item that compounds fastest. Most platforms charge per email verified. At meaningful scale it becomes its own significant cost.
We deploy Reacher — the same open-source engine that underpins several commercial verification products — on dedicated infrastructure. One client we deployed for is running 250,000 to 350,000 verifications per day on a server they own. The per-verification cost is effectively zero. The only cost is the server.
Reacher validates addresses in real time against live mail servers without actually sending an email. Cleaner lists, better deliverability, no per-verification billing — regardless of volume.
What the numbers actually look like
A 30-person team with around 15,000 contacts, sending 100,000 emails per month and verifying 50,000 addresses, before and after migration to self-hosted email infrastructure:
| SaaS stack | Self-hosted | |
|---|---|---|
| Business email (30 seats) | Google Workspace Standard — $420/mo | Mailcow on owned VPS |
| Campaign sending (100k/mo) | Mailchimp Standard — ~$200/mo | Listmonk + Amazon SES — $10/mo |
| Email verification (50k/mo) | ZeroBounce or NeverBounce — $250–300/mo | Self-hosted Reacher — $0 |
| Server | — | $40–60/mo |
| Monthly total | ~$870–920 | $50–70 |
The infrastructure cost doesn't move when the team grows from 30 to 50. It doesn't move when the list doubles. The server changes only if resource requirements meaningfully change — which, in practice, is rare.
Why this isn't a deploy-and-walk-away setup
Email has one property that makes it different from most other infrastructure: deliverability is fragile and slow to recover once it's damaged.
A missed update, an unmonitored spike in bounce rate, a feedback loop complaint nobody saw, an SSL certificate quietly expiring — any of these can affect inbox placement for weeks. By the time someone notices the open rate dropping, the damage is several campaigns deep.
This is why our deployments include ongoing management as the default option. Updates, bounce monitoring, spam complaint tracking, backups, and incident response handled by someone who already knows the client's setup. It's the difference between owning email infrastructure and inheriting a second job.
If this sounds familiar
If you're paying per seat for Google Workspace, per subscriber for Mailchimp, and per check for verification — and the totals are climbing every quarter — the free assessment covers whether the numbers work for your specific situation. Honest answer either way; if self-hosting doesn't make sense at your volume, we'll say so.
If you'd rather see what a migration actually looks like before talking, the rootpath.studio case study walks through one start to finish.
The three invoices don't have to be three invoices. Most of the time they shouldn't be.