Should I use dedicated IPs at 500K emails/month?

Consider it. At 17K emails/day average, you have enough volume to warm and maintain one dedicated IP for marketing. Keep transactional on shared IPs (lower volume, don't want to waste dedicated IP capacity on receipts). Not mandatory—shared still works if your ESP's pools are well-managed.

What ESP tier do I need for 500K/month?

You've outgrown basic tiers. SendGrid Pro ($90-200), Mailgun Scale ($165), or multiple SES quotas. At this level, negotiate—ESPs offer flexibility for committed volume. Don't pay list prices without asking.

How do I split 500K between transactional and marketing?

Typical split: 20-30% transactional (100-150K), 70-80% marketing (350-400K). Route transactional through Postmark or dedicated SES stream. Route marketing through SendGrid/Mailgun with dedicated IP option. Never mix them on the same sending infrastructure.

What monitoring do I need at 500K/month?

Google Postmaster Tools (required), ESP dashboards (daily review), feedback loops with major ISPs, and consider quarterly inbox placement testing with GlockApps or similar. At this volume, you generate enough data for meaningful insights.

Is self-hosted email viable at 500K/month?

Borderline. Self-hosted saves money at this volume ($50-100/month vs $200-400), but requires technical expertise and ongoing maintenance. Consider hybrid: self-hosted for non-critical marketing, ESP for transactional reliability.

Email Infrastructure for 500K Emails/Month

The 500K/Month Profile

At 500K emails/month, you're solidly in mid-market territory:

~17,000 emails/day average
Enough volume to consider dedicated IPs
ESP pricing becomes negotiable
Deliverability issues have real business impact
Self-hosted starts making financial sense

This volume demands more deliberate infrastructure than "just pick an ESP."

Architecture Design

Recommended Structure

yourdomain.com
├── Transactional (100K/month)
│   ├── Domain: mail.yourdomain.com
│   ├── ESP: Postmark or dedicated SES stream
│   └── IP: Shared (reliable high-reputation pools)
│
└── Marketing (400K/month)
    ├── Domain: marketing.yourdomain.com
    ├── ESP: SendGrid/Mailgun with dedicated IP
    └── IP: 1 dedicated IP (or premium shared pool)

Why This Split

Transactional on shared IPs:

Volume (100K/month) doesn't justify dedicated IP warming
Postmark's shared pools have excellent reputation
Password resets and receipts need reliability over control

Marketing on dedicated IP:

400K/month = 13K/day average, enough to maintain IP reputation
You control your reputation (not mixed with other senders)
Marketing campaigns are more variable, benefit from isolation

Practitioner note: At 500K/month, I often see clients on shared IPs who could benefit from dedicated but don't want to deal with warmup. The compromise: stay on shared but use your ESP's premium shared pools if available. SendGrid and Mailgun offer "high-reputation" shared pools for higher-tier customers.

Dedicated IP Considerations

When to Use Dedicated

Yes, dedicated IP:

You send 15K+ emails/day consistently
Your sending schedule is predictable (daily or every other day)
You have clean lists and good engagement
You can absorb the warmup period (4-6 weeks)

No, stay shared:

Sending is inconsistent (weekly blasts, seasonal)
You're not confident in list quality
You can't wait for warmup
ESP's shared pools are performing well

Warmup at 500K/Month Scale

If you add a dedicated IP:

Week	Daily Volume	Notes
1	500-2,000	Most engaged recipients only
2	2,000-5,000	Expand to recent engagers
3	5,000-10,000	Full engaged segment
4	10,000-15,000	Approaching target volume
5+	13,000-17,000	Normal operations

Don't rush. A poorly warmed IP causes months of problems.

ESP Strategy

Option 1: Single ESP with Separation

Use one ESP but separate streams:

SendGrid Pro
├── IP Pool: Transactional (shared)
├── IP Pool: Marketing (1 dedicated IP)
└── Separate API keys per stream

Pros: Single relationship, unified reporting Cons: Single point of failure, less specialization

Option 2: Specialized ESPs

Split by strength:

Transactional: Postmark ($150/month for 150K)
Marketing: Mailgun Scale ($165/month for 350K)

Pros: Best tool for each job, ESP redundancy Cons: More vendors, split reporting

Option 3: SES + ESP Hybrid

Use Amazon SES for marketing volume, ESP for transactional reliability:

Transactional: Postmark (100K, $100/month)
Marketing: Amazon SES (400K, $40/month)

Pros: Lowest cost ($140/month total) Cons: SES requires more management, less support

Pricing Reality

At 500K/month, negotiate:

ESP	List Price	Negotiated
SendGrid Pro	$150-250	$120-180
Mailgun Scale	$165	$140-150
Postmark	$175-225	Usually firm
Amazon SES	$50	No negotiation (already cheap)

Call sales and ask for volume discounts. Most ESPs expect negotiation at this level.

Monitoring Infrastructure

Required: Google Postmaster Tools

At 500K/month, you generate enough Gmail volume for useful data:

Domain reputation (High/Medium/Low/Bad)
IP reputation per sending IP
Spam rate trends
Authentication pass rates

Check weekly minimum, daily during campaigns.

Required: ESP Analytics

Daily review of:

Delivery rate by campaign/stream
Bounce rate (hard vs soft)
Complaint rate
Unsubscribe rate

Set alerts for thresholds:

Bounce rate >3%: Warning
Bounce rate >5%: Critical
Complaint rate >0.1%: Warning
Complaint rate >0.2%: Critical

Recommended: Feedback Loops

Gmail (via Postmaster Tools)
Microsoft (via SNDS and Junk Mail Reporting)
Yahoo (via Complaint Feedback Loop)

Process complaints automatically to remove complainers from lists.

Optional: Inbox Placement Testing

Quarterly testing with GlockApps or similar:

See where mail lands across providers
Identify specific inbox vs spam issues
Track changes over time

Cost: $50-100/month for useful testing volume.

Practitioner note: Inbox placement testing becomes valuable at 500K/month because you have enough volume to care, and enough budget to afford it. Below this, rely on open rates as a proxy for inbox placement.

List Management at Scale

Hygiene Automation

At 500K contacts, manual hygiene doesn't scale:

Hard bounce → Immediate suppression
Soft bounce (3x) → Suppression
Complaint → Immediate suppression + investigation
No opens (6 months) → Re-engagement campaign
No opens (9 months) → Move to sunset segment
No opens (12 months) → Remove from active list

Segmentation Strategy

Segment	Engagement	Send Frequency
Hot	Opened/clicked last 30 days	Multiple per week
Warm	Opened/clicked last 90 days	Weekly
Cool	Opened/clicked last 180 days	Bi-weekly
Cold	No activity 180+ days	Monthly max, re-engage

Send highest-engagement segments first on new campaigns. This builds positive signals before hitting less-engaged recipients.

Cost Analysis

ESP-Only Approach

Component	Provider	Monthly
Transactional (100K)	Postmark	$100
Marketing (400K)	Mailgun Scale	$165
Total		$265

Budget Approach

Component	Provider	Monthly
Transactional (100K)	Amazon SES	$10
Marketing (400K)	Amazon SES	$40
Basic monitoring	Free tier	$0
Total		$50

Requires technical expertise and more hands-on management.

Premium Approach

Component	Provider	Monthly
Transactional (100K)	Postmark	$100
Marketing (400K)	SendGrid Pro	$200
Dedicated IP add-on	SendGrid	$80
Inbox testing	GlockApps	$75
Total		$455

Scaling Considerations

At 500K/month, plan for growth:

Prepare for 1M

Document your infrastructure decisions
Ensure ESP can scale (check tier limits)
Consider second dedicated IP at 750K+
Evaluate self-hosted economics

Warning Signs

Deliverability declining despite good practices
ESP costs becoming significant (>$500/month)
Need features your ESP doesn't offer
Wanting more control over infrastructure

If you're at 500K/month and experiencing deliverability challenges or planning significant growth, schedule a consultation to optimize your infrastructure before issues compound.

Sources

SendGrid: Pro Plan Features
Mailgun: Scale Plan
Google Postmaster Tools: Understanding Your Data
M3AAWG: Sender Best Practices

v1.0 · March 2026