Four of the major SaaS ESPs serve different operational segments. SendGrid, Mailgun, Postmark, and Amazon SES each have distinct positioning, pricing, and operational characteristics. We have worked with customers across all four, often helping them migrate between them or to self-hosted alternatives as their needs evolved.
The comparisons in marketing materials are typically self-serving. The operator perspective is more useful: what each does well, where each falls short, what use cases fit each best.
This post is the honest comparison from an operator who has migrated customers between all four and observed their operational characteristics in production.
The four providers in brief
SendGrid (Twilio SendGrid)
The largest by market share. Acquired by Twilio in 2019. Broad feature set covering transactional, marketing, and infrastructure use cases. Significant marketing automation features.
Positioning: comprehensive solution for many use cases. Strong in B2C marketing.
Mailgun (Sinch)
Developer-focused historically. Acquired by Sinch in 2020. Strong API focus, transactional capabilities, technical credibility.
Positioning: developer-friendly with strong infrastructure capabilities.
Postmark (ActiveCampaign)
Acquired by ActiveCampaign in 2022. Smaller than SendGrid/Mailgun. Strong transactional focus, premium pricing for premium service.
Positioning: transactional-focused with quality emphasis over volume.
Amazon SES
Part of AWS. Different pricing model (volume-based, very low per-message cost). Less feature-rich than dedicated ESPs.
Positioning: infrastructure-level email sending for AWS-integrated operations.
Pricing comparison at various volumes
The pricing landscape as of late 2023:
At 50K monthly messages
- SendGrid: ~€60/month (Essentials plan)
- Mailgun: ~€80/month (Foundation plan)
- Postmark: ~€55/month (Outbound plan)
- Amazon SES: ~€5/month (just messaging costs)
SES is dramatically cheaper at this volume. The premium for dedicated ESPs reflects features beyond just sending.
At 250K monthly messages
- SendGrid: ~€200-€400/month
- Mailgun: ~€225/month
- Postmark: ~€200-€350/month
- Amazon SES: ~€25/month
The dedicated ESPs are 8-15x more expensive than SES at this volume.
At 1M monthly messages
- SendGrid: ~€900-€1,500/month
- Mailgun: ~€850/month
- Postmark: ~€800-€1,200/month
- Amazon SES: ~€100/month
The gap continues widening at higher volumes.
At 5M monthly messages
- SendGrid: ~€3,500-€5,500/month
- Mailgun: ~€3,000/month
- Postmark: ~€2,500-€4,000/month
- Amazon SES: ~€500/month
The dedicated ESPs are 5-10x more expensive than SES at this scale.
At 10M+ monthly messages
- SendGrid: ~€7,000-€12,000/month
- Mailgun: ~€6,000-€10,000/month
- Postmark: ~€4,500-€8,000/month
- Amazon SES: ~€1,000/month
The pricing differences are substantial at scale.
What the pricing reflects
The cost premium for dedicated ESPs over SES reflects significant feature and service differences.
Features bundled with dedicated ESPs
- Email template editors and management
- Subscriber list management
- Marketing automation workflows (varies by provider)
- Analytics and reporting dashboards
- A/B testing capabilities
- Deliverability operations and monitoring
- Customer support
- Documentation and learning resources
What SES does not include
- Template management beyond basic
- List management
- Marketing automation
- Built-in analytics dashboards (CloudWatch integration available)
- Premium customer support
- Deliverability consulting
Operators using SES typically integrate it with other tools to provide the missing features. The integration work is real but enables the cost savings.
Operator-level comparison by category
Deliverability quality
SendGrid: variable depending on the IP pool. Shared IP pools include many senders; quality varies. Dedicated IPs available at premium pricing. Reputation outcomes depend heavily on operator practices.
Mailgun: similar to SendGrid. The shared pool quality depends on which pool. Dedicated IPs available. The technical configuration is more accessible than SendGrid for operators wanting control.
Postmark: strong deliverability due to dedicated focus on transactional. They actively manage their reputation aggressively. Operators get good deliverability with less effort.
Amazon SES: variable. The default pools have mixed reputation. Dedicated IPs available but require explicit configuration. Operators handle most deliverability work themselves.
Operational reliability
SendGrid: significant infrastructure with occasional outages. Status page shows incident history. Reliability is generally good but not exceptional.
Mailgun: similar reliability characteristics to SendGrid. Status page available.
Postmark: reliability is generally very good. Smaller infrastructure but well-managed.
Amazon SES: AWS-level reliability typically. The infrastructure benefits from AWS’s broader operational excellence. Occasional service-specific issues.
Customer support quality
SendGrid: variable. Standard support is bounded; premium support available at premium pricing. Response quality varies.
Mailgun: similar to SendGrid. Documentation is generally good. Direct support varies.
Postmark: notably good customer support. The premium pricing reflects in support quality.
Amazon SES: AWS support model applies. Basic support included; premium support available. AWS support is generally responsive but generic; email-specific expertise varies.
Developer experience
SendGrid: extensive SDKs and APIs. Documentation is comprehensive. Integration patterns are well-established.
Mailgun: developer-focused from the start. Documentation is excellent. APIs are well-designed.
Postmark: clean APIs and good documentation. Smaller surface area but well-executed.
Amazon SES: AWS SDK integration. Documentation is AWS-style (comprehensive but sometimes complex). Integration follows AWS patterns.
Marketing features
SendGrid: most comprehensive marketing automation among the four. Strong B2C marketing capabilities.
Mailgun: less marketing-focused than SendGrid. Functional but not the strength.
Postmark: explicitly transactional-focused. Marketing features are minimal.
Amazon SES: essentially no marketing features. Infrastructure-only.
Transactional features
SendGrid: good transactional capabilities alongside marketing features. The dual focus can create complexity.
Mailgun: strong transactional focus. APIs designed for high-volume transactional sending.
Postmark: explicitly transactional-focused. The product is optimized for this use case.
Amazon SES: appropriate for high-volume transactional. The simplicity matches the use case.
Content policy enforcement
SendGrid: relatively strict content policy enforcement. Suspensions for policy violations occur. Appeals processes exist but are slow.
Mailgun: similar policy enforcement to SendGrid. Suspensions occur.
Postmark: very strict content policy. They reject use cases that other ESPs would accept. The strictness produces their high deliverability reputation.
Amazon SES: bounded policy enforcement. The infrastructure approach allows more diverse use cases. Specific abuse cases trigger enforcement.
Account suspension risk
SendGrid: real suspension risk for senders in marginal content categories. Historical pattern of suspensions during enforcement waves.
Mailgun: similar suspension risk. Their suspensions can come with limited notice.
Postmark: high suspension risk for senders not in their preferred use case. They actively reject mismatched customers.
Amazon SES: lower suspension risk for compliant operations. The infrastructure approach is more permissive than dedicated ESPs.
Which ESP fits which use case
Based on the characteristics:
B2C marketing operations
Best fit: SendGrid for comprehensive marketing capabilities. Mailgun as alternative with developer focus.
Less fit: Postmark (transactional focus). Amazon SES (lacks marketing features).
B2B SaaS transactional
Best fit: Postmark (transactional focus with quality). Mailgun (developer-friendly transactional).
Less fit: SendGrid (overkill for transactional-only). Amazon SES (works but requires more integration).
High-volume transactional (above 5M monthly)
Best fit: Amazon SES (best economics at scale). Mailgun (good capability at lower cost than SendGrid).
Less fit: SendGrid and Postmark (pricing becomes uncompetitive).
Newsletter publishers (B2C marketing-adjacent)
Best fit: SendGrid (comprehensive features). Mailgun (alternative).
Less fit: Postmark (suspension risk for marketing content). Amazon SES (lacks newsletter features).
Marketing agencies (multi-client)
Best fit: SendGrid or Mailgun (multi-account capabilities, sub-account features).
Less fit: Postmark (transactional focus). Amazon SES (multi-account through AWS organization is possible but complex).
Developer-focused operations
Best fit: Mailgun (best developer experience). Postmark (clean APIs). Amazon SES (if AWS-aligned).
Less fit: SendGrid (more complex APIs).
Cost-sensitive high-volume
Best fit: Amazon SES.
Less fit: All three dedicated ESPs are expensive at scale.
Content category in mainstream restrictions
Best fit: None of the four. Self-hosted alternatives.
Less fit: All four have content restrictions that affect specific categories.
Cryptocurrency or adult content
Best fit: None of the four have explicit policies welcoming these segments. Self-hosted required.
Cold email operations
Best fit: None. SendGrid and Mailgun have policies against cold email. Postmark explicitly prohibits it. Amazon SES allows but reputation outcomes are typically poor.
Cold email at scale requires either multi-mailbox approaches (covered in another post) or specialized providers.
What we have seen with customers
Across customers who have used various combinations:
SendGrid to self-hosted migrations
The most common migration pattern. Customers crossing volume thresholds where SendGrid pricing becomes uncompetitive. The migrations we have done span various customer sizes from 250K monthly to 8M monthly.
Customer satisfaction post-migration: generally positive. Cost savings substantial. Operational complexity higher but manageable with support.
Mailgun to self-hosted migrations
Similar pattern to SendGrid migrations. Customers tend to be more technically sophisticated (matching Mailgun’s developer focus).
The transition is operationally smoother because customers have stronger technical capability.
Postmark migrations
Less common. Postmark’s strict content policy means customers in marginal categories cannot use them in the first place. Customers who fit Postmark’s profile typically stay with Postmark.
When Postmark customers migrate, it is usually due to volume crossing into cost-uncompetitive territory.
Amazon SES migrations
Different pattern. SES customers typically migrate to other approaches because they want features SES does not provide (better template management, better deliverability operations, customer-facing dashboards).
The migrations are often to managed self-hosted infrastructure rather than to other SaaS ESPs.
Multi-ESP setups
Some customers use multiple ESPs for different purposes. Transactional through Postmark, marketing through SendGrid, very high-volume through SES. The complexity is manageable for sophisticated operations.
The multi-ESP approach provides redundancy and use-case alignment but increases operational complexity.
The self-hosted alternative
Across all four ESPs, the self-hosted alternative becomes relevant at sufficient scale.
Scale threshold
The crossover point varies by operator but typically:
- Below 200K monthly: SaaS ESP usually wins on operational simplicity
- 200K-500K monthly: depends on specific operator factors
- Above 500K monthly: self-hosted often wins on economics
- Above 2M monthly: self-hosted typically substantially wins
Strategic considerations
Beyond raw economics:
- Content category restrictions: self-hosted bypasses ESP restrictions
- Geographic considerations: self-hosted allows jurisdictional flexibility
- Strategic positioning: self-hosted aligns with privacy-aligned positioning
- Operational control: self-hosted provides direct control
What the self-hosted alternative provides
- PowerMTA or Postal for MTA layer
- MailWizz or Acelle Mail for platform layer
- Managed services from providers like us
- Dedicated IPs and infrastructure
- Custom configuration and operations
The combination produces operations that can match or exceed dedicated ESP capability at lower cost at scale.
What the self-hosted alternative requires
- Initial setup investment
- Ongoing operational discipline
- Technical capability or managed services relationship
- Patience with the operational complexity
The investment is justified for operations meeting the scale and capability requirements.
What we typically recommend
For customers asking us about ESP selection:
For operations that need it now
Use SaaS ESP. SendGrid for marketing-heavy, Postmark for transactional-heavy, Mailgun for developer-heavy, SES for cost-sensitive high-volume.
The SaaS ESPs are operationally simpler. For operations that need to start sending mail immediately, the SaaS approach is faster to set up.
For operations expecting significant growth
Plan for the migration. SaaS ESP works at lower volumes; self-hosted works at higher volumes. The transition is bounded but real.
Customers who plan ahead for the migration handle it more smoothly than customers who scramble during their crossover point.
For operations with strategic considerations
Consider self-hosted from the start. Content category restrictions, strategic positioning, jurisdictional considerations all favor self-hosted.
The initial investment in self-hosted is bounded; the strategic alignment is significant.
For operations testing the market
SaaS ESP. The risk of investing in self-hosted before validating product-market fit is real.
Once the operation is validated, the migration to self-hosted can happen if economics justify.
For operations comfortable with technical complexity
Either path can work. Self-hosted provides more control; SaaS ESP provides more convenience.
The choice depends on what the team values more.
What we recommend against
Several patterns we see that produce poor outcomes.
Provider hopping
Customers who switch between SaaS ESPs frequently produce poor reputation outcomes. Each new provider requires IP warmup. The accumulated reputation does not transfer.
Choose a provider and commit to it for at least 12-18 months unless specific issues force change.
Self-hosted without capability
Customers attempting self-hosted without operational capability or external support produce poor outcomes. The infrastructure does not maintain itself.
Either build the capability or use managed services. Going alone without capability fails predictably.
SaaS ESP at scale beyond reasonable economics
Customers paying €10K+/month to SaaS ESPs when self-hosted would cost €1-2K/month are leaving significant money on the table. The savings could fund operational priorities.
The cost-benefit should drive decisions. Inertia is not a good reason to overpay.
Multi-ESP without operational discipline
Multi-ESP setups can work but require operational discipline. Customers using multiple ESPs without managing them properly produce more issues than benefits.
Use multi-ESP only with explicit operational plan.
The longer-term landscape
Looking at how this might evolve:
Continued ESP consolidation
ESP acquisitions have been frequent. SendGrid (Twilio), Mailgun (Sinch), Postmark (ActiveCampaign), Mailjet (Sinch), various smaller acquisitions. The consolidation may continue.
The implications for operators: the ESP they choose today may be owned by different parent company in a few years. Strategic decisions about ESP dependence should consider this.
Self-hosted ecosystem maturation
The self-hosted ecosystem continues maturing. Better tools, better managed services, more operator expertise. The barrier to self-hosted operation continues lowering.
The trajectory favors self-hosted alternatives over time for operations at meaningful scale.
Receiver-side enforcement tightening
Major receivers continue tightening requirements. The impact varies by ESP. ESPs with stronger compliance discipline are positioned better for tightening enforcement.
The implications for operators: compliance discipline matters. ESPs that produce compliant outcomes serve operators better than ESPs whose compliance is marginal.
Pricing model evolution
ESP pricing continues evolving. The headline pricing comparisons may shift over time. Operators should periodically reassess ESP economics.
Specialized providers
Smaller providers focused on specific niches continue emerging. Privacy-focused ESPs, regional ESPs, vertical-specific ESPs. The choices for operators expand.
What we tell customers about the comparison
For customers asking us about ESP selection:
The four major providers serve different use cases. Match your use case to the appropriate provider.
Pricing matters. The differences are significant at scale.
Operational fit matters. The provider that aligns with your technical capability and use case produces better outcomes.
The choice is not forever. Migrations between providers are possible. Choose the right provider for now; revisit periodically.
Self-hosted alternatives exist. At sufficient scale, self-hosted becomes economically and operationally competitive.
We can help with various paths. SaaS ESP setup, migration between ESPs, migration to self-hosted, multi-ESP coordination.
The honest assessment of your situation produces better decisions than provider-specific marketing materials.
The customer profile we serve
Our customers come from various ESP backgrounds:
- Approximately 35% migrated from SendGrid
- Approximately 20% migrated from Mailgun
- Approximately 10% migrated from Postmark
- Approximately 15% migrated from Amazon SES
- Approximately 20% started directly with us (no prior ESP)
The migration patterns reflect ESP characteristics and our positioning. SendGrid migrations are most common because SendGrid is largest. Mailgun migrations are second because Mailgun customers are technically sophisticated. Postmark migrations are less common because Postmark’s profile aligns with customers who stay with them.
Each migration has informed our understanding of the source ESP and the migration process.
The honest summary
The four major SaaS ESPs each serve operational segments. The honest assessment of your specific situation should drive the choice rather than marketing influence.
SendGrid for comprehensive marketing-heavy operations. Mailgun for developer-focused operations. Postmark for premium transactional with willingness to pay. Amazon SES for cost-sensitive high-volume with AWS alignment.
Self-hosted alternatives become relevant at scale and for specific strategic considerations.
The choice is bounded. The migration between providers is possible but should not be casual. The strategic considerations beyond raw pricing matter.
For our customer base, the mix reflects various paths and use cases. We help customers across all paths. The conversations we have with customers about provider selection produce better-aligned decisions than customers make independently based on marketing materials.
The four providers will continue evolving. The comparison will continue being relevant. The general patterns described in this post will continue applying with specific updates as providers change.
For operators considering ESP selection today: honest assessment of your situation, comparison against the operator-level characteristics, and willingness to revisit the decision as your operation evolves produce better outcomes than provider-loyalty or marketing-driven choices.