Dedicated vs Shared IP Address

What this is for#

Your sending IP's reputation directly determines whether your emails land in the inbox or spam folder. The choice between a dedicated IP (yours alone) and a shared IP (in a pool with other senders) sits at the center of every deliverability conversation.

This guide walks the trade-offs by volume tier, the warmup commitment a dedicated IP demands, the failure modes of each, and the hybrid pattern that high-volume senders actually use.

The core comparison#

When to choose dedicated#

Dedicated IP makes sense when all three of these are true:

1. Volume: consistently send >50,000 emails/month (some say >100k). Below that, an under-utilized IP is a "low-volume warning" signal to receivers — they treat it as suspicious.
2. Steady cadence: you send regularly (multiple campaigns per week). A dedicated IP that sits idle for weeks loses warmth fast.
3. Operational capacity: you can monitor reputation (Postmaster Tools, SNDS), respond to reputation dips, and execute warmup if you migrate.

If any of those is false, dedicated will cost you reputation, not help it.

When shared is the right answer#

Shared makes sense when:

• You send <50k/month — your volume can't keep a dedicated IP "warm"
• You're new — you don't have a track record to leverage
• Your provider's pool is reputable — Amazon SES + SendGrid + Mailgun + Postmark all run well-managed shared pools where the pool's overall reputation is good
• Your sending is bursty — a dedicated IP can't tolerate "0 sends Mon-Fri, 50k blast on Sunday"; shared pools smooth this out

The 50k/month threshold#

Why this specific number? Three reasons converge:

1. Reputation building: receivers (Gmail Postmaster, Microsoft SNDS) need sustained volume to compute a reputation score. Below ~1500/day = 45k/month, you're in the "unknown sender" bucket where every send is judged in isolation, not in pattern.
2. Warmup math: a 6-week warmup ramps from ~200/day to 25k/day. To "earn" the warmup investment, you need to sustain that 25k/day after — that's ~750k/month. The 50k/month threshold is the lower bound where warmup ROI starts.
3. Cost math: dedicated IPs typically cost $30-100/month extra. At 50k/month that's $0.06-$0.20 per 1000 sends extra cost. Below 50k it's a tax; above 500k it's negligible.

The hybrid pattern (high-volume senders)#

At enterprise scale (millions of sends/month), the right answer is neither pure-dedicated nor pure-shared — it's a multi-IP architecture:

• Transactional traffic (password resets, order confirmations) → 1 dedicated IP with high reputation
• High-priority marketing (welcome emails, behavioral triggers) → 2-3 dedicated IPs in rotation
• Bulk newsletter → 2-4 dedicated IPs in rotation
• Re-engagement / risky content → shared pool OR a separate "tier-2" dedicated IP that can be sacrificed if it gets blocked

Why segment: an automated spam complaint on your re-engagement campaign shouldn't damage the IP that delivers your password resets. Separation is reputation insurance.

For implementation, see Multi-Server Rotation Pattern.

Warmup commitment for dedicated#

A new dedicated IP starts at zero reputation with every major receiver (Gmail, Outlook, Yahoo). They treat unknown IPs as suspicious. To build trust:

Monitor at every step:

• Bounce rate < 2% per send
• Spam complaint rate < 0.1% per send
• Gmail Postmaster reputation: "High" or "Medium" (never "Low" or "Bad")
• Microsoft SNDS color: Green (Yellow is acceptable, Red is stop-immediately)

Any spike → halt warmup, investigate, resume only after the cause is fixed. See IP Warmup Schedule for the full per-week protocol.

Recovery from a reputation hit#

On dedicated IP: weeks-to-months. You'll see open rates drop, more sends go to spam folder, possibly some receivers temporarily blocking your IP. Recovery requires reducing volume + sending only to engaged subscribers + waiting for receivers to re-evaluate (Gmail re-scores reputation every 24-48h based on recent behavior).

On shared IP: depends on the provider. SES, SendGrid, Mailgun rotate misbehaving senders into "warming pools" or off the main pool — your reputation recovers as the bad neighbor is removed. You feel a temporary dip, then it self-heals.

How to provision a dedicated IP#

Amazon SES#

SES Console → Configuration → Dedicated IPs → Lease dedicated IP

Costs ~$24.95/month per IP. Assign to a "dedicated IP pool" → assign your sending identity to that pool. New IP starts warm (SES does an auto-warmup over 45 days).

SendGrid#

SendGrid Dashboard → Settings → IP Addresses → Add IP

~$80-100/month. SendGrid manages warmup automatically over 30-45 days.

Mailgun#

Mailgun Control Panel → Sending → Dedicated IPs → Request

~$59/month. Mailgun's warmup tool is built-in.

Your own SMTP / on-prem#

Provision via your hosting provider (DigitalOcean Floating IPs, AWS Elastic IPs, Hetzner additional IPs). You manage warmup manually. Most flexibility, most operator burden.

Failure modes to be aware of#

Dedicated IP failures#

• Underutilization: IP sits idle for 1-2 weeks → reputation decays → first big campaign sees deliverability drop
• Sudden volume spike: going from 5k/day to 50k/day in a single day looks like spam-bot behavior to receivers → temporary throttling
• List quality drop: importing a stale or poorly-cleaned list onto a warmed dedicated IP burns the reputation you built
• Single-IP outage: if your dedicated IP gets blocked, you have no fallback — sending stops

Shared IP failures#

• Neighbor abuse: another sender on your shared pool has a deliverability incident → pool reputation drops → your sends affected
• Provider migration: sending provider quietly moves you to a different shared pool with worse reputation — opaque to you
• No visibility: you can't see why your shared-pool deliverability changed; only the provider has the data
• Volume cap: shared pools throttle individual senders during high-load periods → your sends queue when you most want them out

Common issues#

FAQ#

Can I have BOTH dedicated and shared on the same account? Yes — most providers support this. Use dedicated for primary marketing + transactional, shared as fallback or for low-priority sends.

Does AcelleMail itself care which I use? No — Acelle just calls the sending provider's API. The dedicated-vs-shared decision happens at the provider level.

Cost-effective alternative: multiple smaller dedicated IPs? Yes — at 250k+/month, splitting into 2-3 smaller dedicated IPs ($75/mo total) gives you rotation + reputation insurance for less than 1 large dedicated. See Multi-Server Rotation Pattern.

What's a "warming pool"? Some providers (notably SES) maintain pools specifically for senders ramping up — new dedicated IPs join the warming pool, gradually graduate to the main pool. Reduces warmup burden on operators.

IP reputation across providers — does it transfer? No. An IP's reputation is per-IP, not per-sender. If you migrate to a new IP (even at the same provider), you start over.

My audience is mostly mobile. Does that change the decision? No — receivers grade based on send/engagement patterns, not on recipient device. Mobile vs desktop is the rendering concern, not the deliverability concern.

Geographic dedicated IPs (one per region)? Useful at very large scale (multi-region campaigns) but rarely cost-justified below 10M/month per region. Most senders are fine with a single IP serving global audiences.

Related articles#

• IP Warmup Schedule for New Sending Servers — the warmup protocol if you go dedicated
• IP Warmup by ISP — per-receiver warmup nuances
• Multi-Server Rotation Pattern — for the hybrid multi-IP architecture
• Sender Reputation Monitoring — Postmaster + SNDS + your own metrics
• Gmail Postmaster Tools Walkthrough — reputation visibility for Gmail
• Microsoft SNDS Walkthrough — reputation visibility for Outlook
• Configuring Amazon SES with AcelleMail — SES dedicated IP setup
• SES Sending Limits Cookbook — provider-side quotas
• Deliverability Incident Runbook — what to do when reputation tanks
• Reverse DNS and PTR Records (in DNS Setup guide) — PTR is per-IP; dedicated IPs need their own PTR