How to evaluate outbound-email infrastructure.
Every cold email provider quotes a price per inbox. The differences that decide whether your campaigns survive are structural: density, isolation, ownership, warmup, monitoring, and what happens when a domain degrades.
Five ways to buy sending infrastructure.
These are models, not vendors — most providers fit one of them cleanly. The right question isn't which logo, it's which operating assumptions you're buying into.
| Direct Workspace / 365 | Ultra-cheap reseller | High-density Azure | Custom SMTP / dedicated | Mailrun | |
|---|---|---|---|---|---|
| Infrastructure | Google or Microsoft seats administered by the buyer | Workspace, Microsoft, or SMTP inventory; licensing and tenancy vary | Microsoft / Azure infrastructure optimized for higher mailbox counts | Self-managed servers, MTAs, and IP infrastructure | Licensed Microsoft / Azure infrastructure operated at controlled density |
| Inboxes per domain | Set by the buyer; no outbound-specific guardrail by default | Provider-defined; ask for the enforced maximum | Provider-defined; verify density and isolation | Configuration-dependent | 10 max, typically lower |
| Reputation isolation | One tenant — often your real company domain | Varies by provider and plan; ask how tenants are separated | More inboxes share each domain’s reputation; ask how load is isolated | Your IPs carry everything | Low density, domain-level isolation |
| Domain ownership | Yours | Terms vary; confirm ownership and transfer rights | Terms vary; confirm before you build reputation on them | Yours | Managed & BYOD: yours. Pre-warmed: subscription inventory. |
| Time to mailbox access | Days of DIY setup | Fast | Fast | Weeks of build-out | Hours |
| Warmup | Do it yourself | Varies; confirm whether warmup is included, required, or customer-managed | Varies; confirm the recommended ramp | Do it yourself | Two-week floor, controlled ramp; pre-warmed available |
| Monitoring | Not included by the mailbox provider by default | Varies by provider; ask what signals you can actually see | Varies by provider | Customer-built | Bounce, complaint, and domain-health signals watched continuously |
| When a domain degrades | You notice, you fix | Replacement and remediation policies vary; ask what happens to the domain itself | Replacement and remediation policies vary | You rebuild | Bench, repair, or replace — while the pool absorbs the volume |
| Sequencer handoff | Manual | Varies; confirm the export format | Varies | Manual | Direct API for major sequencers, CSV for everything else |
| Pricing structure | Per seat, priced like employees | Per-inbox or package pricing; verify commitments, add-ons, and domain costs | Per inbox | Infrastructure plus your time | Flat monthly by capacity, public prices |
| You stay responsible for | Everything | Campaigns, plus whatever the provider doesn’t operate — ask precisely what that is | Campaigns; verify how much of the operating falls to you | Everything | Campaigns, copy, and lists — we run the infrastructure |
Descriptions reflect how each model typically operates, not any specific vendor. Individual providers vary — which is exactly why the evaluation questions below matter more than the category.
Ask these before you buy from anyone — including us.
Build the sending plan your next launch deserves.
Public pricing, hard density caps, and a bench-and-repair workflow you can see. If another model fits you better, we'll tell you that on the call.
Mailbox access in hours. Sending volume ramps with discipline.
