# The email marketing stack in 2026 — tools, architecture, and tradeoffs

Modern email-marketing systems are stacks of three to five interchangeable layers. Picking the right pieces — and deciding which layers you own versus rent — drives both cost-per-send and deliverability ceiling.

## The five layers

1. **Marketing app (ESP).** List management, campaign authoring, automation flows, segmentation, analytics. The user-facing surface.
2. **Sending pipeline.** The MTA + IP reputation. Most teams hand this to a managed relay (Amazon SES, SendGrid, Mailgun, Postmark, SparkPost).
3. **Deliverability tooling.** SPF / DKIM / DMARC records, FBL processing, postmaster dashboards, seed-list testing.
4. **Analytics + attribution.** UTM tagging, link tracking, downstream revenue tie-back.
5. **Consent + identity.** Opt-in source, double opt-in, preference center, suppression list, GDPR + CAN-SPAM.

The big question: which layers do you OWN vs RENT? Big-SaaS rents all five. Self-hosted operators own 1, 3, 4, 5 — and rent layer 2 from a cloud SMTP. That split — own the data, rent the sending muscle — is the dominant 2026 pattern.

## Three reference architectures

- **A. All-in-one SaaS** (Mailchimp / Brevo / Klaviyo). One signup, all five layers. Right for under 5K contacts with low engineering capacity.
- **B. Marketing app + cloud SMTP** (AcelleMail / Listmonk / Mautic on a $5–$20 VPS, sending through SES). Owner controls the data + campaign UX; relay handles IP reputation. Cost-per-send approaches $0.10 per 1,000 emails at scale.
- **C. Transactional API + custom marketing tooling** (SendGrid API / Postmark / Mailgun + internal app). Used by SaaS products where most email is transactional.

## Buy, build, or compose?

- **Buy (A)** — under 5K contacts, low engineering capacity, ship-speed matters more than cost.
- **Compose (B)** — over 5K contacts, can stand up a VPS, want one-time license + linear cost-per-send. AcelleMail + SES is the canonical example.
- **Build** — only if your business IS the marketing platform. Otherwise composing existing pieces is dramatically faster.

The common 2026 mistake: staying on Architecture A past the cost-curve inflection. Architecture B at 5K–50K contacts pays back the migration time in under a year.

## Tool inventory by layer

- **Marketing apps (open-source).** AcelleMail, Listmonk, Mautic, Postal, Sendy.
- **Marketing apps (SaaS).** Brevo, Campaign Monitor, ConvertKit / Kit, Klaviyo, Mailchimp, MailerLite, Sendinblue.
- **Cloud SMTP relays.** Amazon SES, Mailgun, MailerSend, Postmark, SendGrid, SparkPost.
- **Self-managed MTAs.** exim, Haraka, Postal, Postfix.
- **Seed-list testers.** GlockApps, Inbox-Insight, Mail-Tester.
- **Postmaster dashboards.** Apple iCloud Postmaster, Gmail Postmaster Tools, Microsoft SNDS, Yahoo SmartNetwork.

## Where to go next

- [Self-hosted email-marketing pillar](https://acellemail.com/guide-self-hosted.md)
- [Cost-savings deep dive](https://acellemail.com/guide-cost-savings.md)
- [Deliverability fundamentals](https://acellemail.com/guide-deliverability.md)
- [Side-by-side: AcelleMail vs Mailchimp](https://acellemail.com/vs-mailchimp.md)