Most people treat email as one thing. There are at least seven distinct types, and each one has different infrastructure requirements, different risk profiles, and different consequences when something goes wrong.
Run them all through the same domain and you'll end up sending cold outreach from your main business domain, watching your transactional emails land in spam, and wondering why your newsletter open rates cratered. The cause is almost always infrastructure, not copy.
Here's the full taxonomy and what each type actually needs.
Transactional Emails
Transactional emails are triggered by a user action: a purchase confirmation, a password reset, a shipping update, an invoice. The recipient completed an action that makes them expected, so they arrive with high engagement rates almost by default.
They also carry the highest stakes when they fail. A password reset that lands in spam means a locked-out customer. An order confirmation that doesn't arrive means a support ticket.
Infrastructure requirements: a dedicated sending domain separate from everything else, a dedicated SMTP server or transactional ESP (SendGrid, Postmark, Mailgun), and SPF/DKIM/DMARC configured correctly. Sender reputation matters here, but because engagement rates are high on transactional mail, reputation tends to stay healthy as long as bounce rates stay under 2%.
The rule: never mix transactional sending with marketing or outreach sending. Shared infrastructure means shared consequences. One bad campaign tanks the deliverability of your account confirmations.
Marketing and Newsletter Emails
Marketing emails go to opted-in subscribers: product updates, promotional campaigns, weekly newsletters, new feature announcements. Everyone on the list asked to be there.
This type carries more regulatory weight than others (CAN-SPAM, GDPR, and CASL all apply directly), so you need proper unsubscribe handling and list hygiene. Sending to stale or unverified lists damages email deliverability fast. Spam complaints spike, sender reputation drops, and you start filtering into spam for engaged subscribers too.
Infrastructure: a dedicated marketing ESP (Mailchimp, Klaviyo, HubSpot) or your own sending infrastructure, on a domain or subdomain separate from your main business domain. Most teams use something like mail.company.com or news.company.com as the sending subdomain to keep reputation isolated from transactional and outreach sending.
Cold Email
Cold email goes to people who haven't given you permission to contact them. It's prospecting: reaching people who fit your ICP but have no prior relationship with you. Executed well, cold outreach is one of the most cost-effective B2B lead generation channels. Executed poorly, it burns domains and damages deliverability across everything else you send.
The infrastructure requirements for cold email are the most demanding of any email type:
- Secondary sending domains: cold outreach generates bounces and complaints at rates that would destroy your main domain's reputation inside a month. Never send from your primary business domain.
- Real provider inboxes: Google Workspace or Microsoft 365 accounts, not cheap SMTP relays. Receiving servers trust mail from established providers far more than mail from generic SMTP infrastructure.
- Email warmup: every new inbox needs a proper email warmup period before campaign sending. New domains and inboxes have no sending history, which means no trust with receiving servers. Warmup builds that history gradually, typically over 2–4 weeks of increasing volume with positive engagement signals.
- Authentication: SPF, DKIM, and DMARC configured correctly on every sending domain. Without authentication, cold emails fail receiving server checks regardless of content quality.
- Active monitoring: tracking bounce rates, spam complaint rates, and inbox placement continuously. A domain going bad needs to be caught early before it's unsalvageable.
The biggest mistake in cold email is using the same domain or inbox for outreach that you use for anything else. When that domain burns, and at volume it eventually will, you want the damage contained to your sending domains, not spreading to your main business.
Warm Outreach
Warm outreach goes to people with some prior connection: a conference conversation, a mutual contact introduction, someone who downloaded a resource, a lapsed customer. There's no established relationship, but the contact has some context for who you are.
Warm outreach carries lower risk than cold email. Lists are smaller, engagement rates are higher, and complaint rates are lower. You can often send it from a dedicated mailbox on your main business domain as long as volume stays modest (under a few hundred per week per inbox) and your authentication is clean.
If you're doing warm outreach at real volume, say re-engaging 10,000 event attendees, treat it like marketing email. Use proper list hygiene, proper unsubscribe handling, and a sending domain or subdomain that keeps the risk isolated.
Internal and Team Emails
Email between employees and across teams. Internal email is the lowest-risk type: your own servers accept it, there's no external spam filtering involved, and delivery is reliable by default.
The main infrastructure consideration is running internal email through proper organization accounts (Google Workspace or Microsoft 365) rather than personal Gmail accounts. For any business with more than a few employees, managed accounts matter for security, access control, and business continuity.
Internal email is a data governance and security concern, not a deliverability one. Email from sensitive departments (legal, finance, HR) should be handled on accounts with appropriate security settings, two-factor authentication enforced, and retention policies configured.
Notification and Automated Emails
Automated emails triggered by system events rather than user actions: account activity alerts, scheduled reports, usage summaries, monitoring notifications. These often come from an SMTP server connected to your application or internal tooling.
From an infrastructure standpoint, automated notification emails need the same clean setup as transactional email: dedicated sending identity, proper authentication, low bounce rates. The failure mode here is reliability, not spam. If your monitoring alerts route through a misconfigured SMTP server, they may not arrive when you actually need them.
Keep automated notification sending separate from marketing and outreach sending. The volume is usually lower and the stakes, when something breaks at 2am, are higher.
Re-engagement Emails
Re-engagement emails go to subscribers or customers who've gone quiet: people who used to open your marketing emails and stopped, or customers who haven't logged in or purchased in several months. The goal is to reactivate them or confirm they want to unsubscribe.
This type is high-risk from a deliverability standpoint. You're emailing people who have already shown disinterest by not engaging. Sending re-engagement campaigns to large cold lists without proper hygiene first generates spam complaints and suppresses sender reputation across your marketing infrastructure.
Verify email addresses before sending re-engagement campaigns. Suppress anyone who hasn't opened in 12+ months unless you're prepared to accept the deliverability hit. Re-engagement is the email type most likely to expose list hygiene problems that were always there but hadn't caused visible damage yet.
Why Mixing Email Types on One Domain Is a Mistake
Each email type generates different engagement signals. Transactional email gets high open rates. Cold outreach gets spam complaints. Marketing email generates unsubscribes. Automated notifications get ignored.
When you run multiple email types through the same domain, all those signals blend together and affect a single domain reputation. A cold outreach campaign that generates spam complaints damages the reputation of the domain your transactional emails send from. A stale marketing list that tanks newsletter deliverability shares a domain with your sales team's warm outreach.
The right architecture keeps these separated:
- Primary domain (
company.com): day-to-day team email, warm outreach at modest volume - Transactional subdomain (
mail.company.comor a separate domain): password resets, purchase confirmations, account alerts - Marketing subdomain (
news.company.comor similar): newsletters, promotional campaigns - Cold outreach domains (multiple, completely separate from
company.com): prospecting campaigns, dedicated inboxes, isolated reputation
One domain burns and the blast radius is contained. You fix it or replace it, and everything else keeps running.
Most teams don't set this up until something goes wrong. At Beanstalk we see it constantly: a company comes in with cold email campaigns running off company.com, transactional emails that started landing in spam six months ago, and nobody connected the dots until a client complained about missing invoices. The fix works, but rebuilding domain reputation takes time.
Cold Outreach Infrastructure: What a Clean Setup Looks Like
Cold outreach is the email type where most teams underinvest in infrastructure and pay for it in deliverability problems. Here's what a clean cold email setup actually requires:
Multiple secondary sending domains, typically 1 domain per 2–3 inboxes as a starting ratio. If one domain gets flagged, you retire it without losing the others. Each domain gets its own SPF, DKIM, and DMARC records, its own warmup period, and its own sending history.
Real provider inboxes (Google Workspace or Microsoft 365) on those domains rather than a generic SMTP server. Receiving mail servers treat mail from established providers more favorably than mail from unknown SMTP infrastructure. The trust differential matters.
A proper email warmup process for every new inbox before it touches any real campaign. This typically takes 2–4 weeks of gradually increasing sends with positive engagement signals. Skip this and you're burning domains faster than you can replace them.
Active monitoring that tracks IP reputation, domain reputation, bounce rates, and spam complaint rates. The cold email infrastructure that keeps running clean catches problems at the first warning sign rather than after the domain is already flagged.
At ScaledMail, we provision and manage this entire stack: Google Workspace and Microsoft 365 inboxes on dedicated sending domains, authentication configured correctly, and warmup handled before you start sending. We monitor continuously. If a domain starts showing warning signs, we flag it before it becomes a deliverability problem that takes weeks to fix.
The infrastructure decisions you make for each email type determine whether any of this actually works. There's no version of this where the setup doesn't matter.
