← Back to all posts

Dedicated IP vs Shared IP: What Your Email Volume Actually Demands

Most "deliverability advice" treats a dedicated IP as a milestone — buy one when you're serious. That's backwards. A dedicated IP before you have the volume to feed it is a reputation liability, not a badge. The math from AWS's own guidance makes it plain.

Why ISPs ignore small dedicated senders

The core problem: mailbox providers (Gmail, Outlook, Yahoo) mostly track IP reputation, and they only build a reputation profile for an IP that sends meaningful volume. Amazon's SES documentation is explicit: ISPs track the reputation of a given IP only if they receive significant volume from it — as a rule of thumb, several hundred emails within a 24-hour period, at least once per month, per ISP you want to build reputation with.

Run that math for a SaaS sending 5,000 emails a month. Split across Gmail, Outlook, Yahoo, and a dozen regional providers, that's maybe 1,500 messages a month to Gmail — far below the threshold where Gmail bothers profiling your IP as anything other than "occasionally quiet." Your dedicated IP stays anonymous. All the deliverability it could offer never materializes, because there's no reputation to leverage.

Meanwhile, you're paying the full price of isolation with none of the benefit.

What a dedicated IP actually costs

Per AWS's published pricing:

• Dedicated IPs – Standard: $24.95 per month, per IP. You manage warm-up, scaling, and pool membership yourself.
• Dedicated IPs – Managed: $15 per month per account plus a per-email fee ($0.08/1,000 at 0–10M emails/mo). SES warms the IPs automatically per-ISP and auto-scales the pool.

That's on top of your sending fees. On the standard tier, one IP for a year is ~$300. For a startup at 50,000 emails a month, you're paying $0.40 per thousand sends in pure IP overhead — before the operational cost.

And the operational cost is real. Warm-up is a gradual daily volume ramp you must schedule and monitor. After warm-up, AWS says you must "maintain a consistent sending pattern" — because an isolated IP with sporadic volume looks worse to filters than an active shared pool IP. Product launches, marketing bursts, seasonal spikes: on a dedicated IP, every anomaly in your sending pattern is now your reputation's problem, with no pool to absorb the variance.

What shared pool gives small senders

A shared pool is a large, actively warmed set of IPs serving thousands of senders at once. The economics flip in your favor:

1. Instant reputation floor. The pool is already warmed and continuously active. Your first send rides existing reputation instead of starting cold.
2. Variance absorption. Your Tuesday product launch spike is noise against the pool's aggregate volume, not a red flag on your own IP.
3. No warm-up schedule. No 4–6 week ramp, no per-ISP pacing calendar, no spreadsheet tracking daily send ceilings.
4. Someone else owns the ops. IP pool health, blocklist remediation, and ISP feedback-loop handling are the provider's job.

The counterargument — "someone else's bad sender pollutes my reputation" — is provider-dependent, not inherent to sharing. Reputable transactional providers police their pools: strict onboarding, spam-trap monitoring, and suspension of abusive senders. A pool where that doesn't happen is a bad provider; it isn't a reason to buy a $25/mo IP you can't feed.

The actual decision rule

The SES guidance effectively defines the crossover: you need, per ISP, several hundred emails within 24 hours at least monthly — sustained and predictable — before an IP-level reputation exists to maintain. Practical thresholds most providers converge on:

• Under ~100,000 emails/month, mixed traffic: shared pool. No exceptions worth arguing about.
• 100,000+/month, consistent daily volume, mostly transactional: start evaluating dedicated IPs. At 200k/month, the $24.95 standard IP is $0.12/1,000 of overhead and you're sending enough per ISP for reputation to form.
• Multi-million/month, separation needs (transactional vs marketing on different IPs): dedicated pools, ideally managed warm-up.

The one legit early-dedicated-IP case: recipients behind strict allowlists (some banks, government gateways) that accept or reject mail by IP rather than reputation. If a handful of corporate recipients blocklists matter more than ISP reputation, ask for dedicated egress — but that's an enterprise edge case, not a growth-stage default.

How this maps to MailAnvil

MailAnvil runs on Amazon SES's shared pool. For the Indonesian startups we build for — 10k–100k emails/month, QRIS billing, IDR pricing — that's the correct engineering choice at every volume they'll have in their first two years: no warm-up debt, no $24.95/mo line item, and pool hygiene handled at the SES level. When your sending volume and consistency genuinely cross the crossover point, the SES dedicated-IP (standard or managed) path is available behind the same API — the migration is configuration, not code.

Shared pool isn't the budget option. It's the reputationally superior option at your volume. Buy the dedicated IP when the math says so — not when your ego does.