Three different customers asked us in the past quarter about building their own ESP business. The pattern is recognizable: they have technical capability, they have observed margins in the mainstream ESP market, they have some volume to start with, and they are interested in the operator-level economics of running an ESP rather than being a customer of one.
The business is viable. The economics work for specific operator profiles. The six-month operational curve catches most new ESP builders by surprise.
This post is the realistic view of building an ESP, based on our observations across customers who have done this successfully and customers who have started and either pivoted or stopped. The capital requirements, the operational curve, the things most teams underestimate, and the segments where ESP building actually makes sense.
Why operators consider building their own ESP
The mainstream ESP market has visible margins. SendGrid, Mailgun, Postmark charge customers €X per thousand messages while the underlying infrastructure cost is fractions of €X. The gross margin is significant.
Operators with technical capability look at this and consider whether they could capture some of that margin by running an ESP themselves. Several specific scenarios:
A digital marketing agency with several clients sends through their own infrastructure rather than paying ESP fees per client. The savings fund the operational investment.
A SaaS founder builds their own ESP as a side business after observing the mainstream ESP economics during their primary business operations.
A privacy-focused operator wants to offer ESP services without the KYC/content restrictions of mainstream providers.
A regional operator targets a specific geographic market underserved by mainstream ESPs.
An existing infrastructure operator (web hosting, VPN, etc.) adds ESP services as a logical extension of existing customer relationships.
Each scenario produces different ESP characteristics and requires different operational decisions. None of them are simple.
The capital requirements
Building an ESP requires capital before revenue. The capital needs cover multiple categories.
Infrastructure capital
Server infrastructure: €1,500-€5,000 to start. The starting infrastructure includes:
- Application servers for the platform layer (MailWizz, Acelle, or custom)
- SMTP servers for the delivery layer (PowerMTA or Postal)
- Database servers for state and analytics
- Monitoring and management infrastructure
Dedicated IPs: €500-€2,000 for initial pool. The cost depends on how many IPs you start with. A small ESP can start with 10-15 IPs; a larger ESP needs 50+ IPs from launch.
Software licensing: variable. MailWizz commercial license is €399-€999 depending on edition. PowerMTA license is €5,000-€15,000 annually if you choose commercial MTA. Postal is free as open-source. Various smaller tools (monitoring, alerting, etc.) range €500-€2,000 annually.
Networking and redundancy: €500-€2,000 if you want redundant network paths, failover capability, geographic distribution.
Total infrastructure capital: €5,000-€20,000 to start, with significantly more if you want enterprise-grade reliability from day one.
Operational capital
Engineering time for setup: 80-160 hours of senior engineering work. At market rates, this is €4,000-€16,000 in time investment.
Documentation and procedures: 40-80 hours to develop the runbooks, customer onboarding documentation, internal procedures. Another €2,000-€8,000 in time investment.
Initial customer acquisition: variable depending on market positioning. Some operators have a captive customer base from existing operations. Others need marketing investment of €5,000-€20,000 to acquire initial customers.
Working capital for revenue lag: 3-6 months of operating cost before revenue covers expenses. Depending on cost structure, this is €15,000-€45,000.
Total operational capital: €25,000-€90,000 in operating capital and time before the ESP is revenue-positive.
Combined capital requirement
For a serious ESP launch with realistic timelines and reasonable quality standards: €30,000-€100,000 of combined infrastructure and operational capital before revenue covers expenses.
The lower end of this range is achievable for operators with existing infrastructure they can repurpose, technical capability that does not need external engineering, and a captive initial customer base.
The higher end of the range is more typical for operators starting from scratch and wanting commercial-grade operations from launch.
Less than €30,000 in committed capital usually produces an undercapitalized launch that struggles in months 3-6 when operational complexity peaks but revenue has not scaled.
The six-month operational curve
The reason ESP builders run into trouble is the operational curve. The work to launch is bounded. The work to operate the ESP six months in is significantly more than most builders anticipate.
Month 1: Launch
The launch month is the most exciting and the least demanding operationally. Infrastructure is set up. First customers are onboarded. Marketing has produced some interest. The team feels good about the progress.
The customer count is small (typically 5-15 customers). The volume is modest. The operational work is mostly infrastructure setup completion and customer onboarding workflow.
Daily time investment: 4-8 hours for the founder/operator. This feels manageable because the customer base is small.
Month 2-3: Growth
Customer count grows (typically 15-50 customers by month 3). Volume scales correspondingly. The infrastructure is mostly stable. Operational patterns start to emerge.
Daily time investment grows: 6-10 hours for the founder/operator. The growth is from customer support, deliverability issues, and ongoing customer acquisition activities.
The challenges that emerge:
- First customer-affecting incidents. Even with good practices, some incidents happen.
- Customer support load grows nonlinearly with customer count.
- Deliverability issues require investigation time.
- IP warmup management for new customers becomes ongoing work.
Month 4: The pressure point
This is where most undercapitalized ESPs face their critical decision. Several things happen in month 4:
Customer count has grown enough (typically 50-100 customers) that the operator can no longer handle everything personally. Either hire help, automate more, or reduce service quality.
Infrastructure has scaled enough that what worked at launch is starting to creak. Database performance issues, queue management edge cases, monitoring gaps all become visible.
Revenue is growing but not yet covering full operational cost. The operator is burning through capital with the expectation that growth will continue.
Operational issues are becoming more complex. Customer ESP usage patterns produce edge cases that did not exist with simpler test traffic.
The choice point: invest more in operational capacity (hire, automate, build better tools) or accept service quality degradation. Operators who invest survive. Operators who accept degradation typically see customer churn that eats into their growth.
Month 5-6: Stabilization or struggle
Operators who invested in month 4 see stabilization in months 5-6. Operations smooth out. Patterns are documented. Customer support load becomes manageable. Revenue starts to cover full operational cost.
Operators who did not invest see continued struggle. Service quality degrades. Customer churn accelerates. Growth stalls. The business either pivots, gets acquired, or shuts down.
The six-month mark is where the operator’s path becomes clear. Continued growth with appropriate operational investment, or stalled growth with operational debt accumulating.
What most teams underestimate
Several specific things consistently catch new ESP builders by surprise.
Customer support load
Mainstream ESPs invest heavily in customer support. SendGrid’s support staff is in the hundreds. Mailgun has significant support infrastructure. The visible product is the platform; the invisible product is the support.
New ESPs assume customer support will be light because their customers are technical. This assumption is wrong. Technical customers have more complex questions, not fewer. They want detailed answers, not brushed-off responses. The support quality expectation is high.
The realistic support load: 30-60 minutes per customer per month on average. With 50 customers, this is 25-50 hours per month of support time. With 200 customers, this is 100-200 hours per month. The support load scales linearly with customer count.
Mainstream ESPs cover this with dedicated support staff. New ESPs typically have the founder doing support, which becomes unsustainable around 50-100 customers.
Deliverability operations
Every customer has deliverability concerns. The new ESP becomes the responsible party for these concerns regardless of whether they technically can fix them all.
Customers expect the ESP to:
- Monitor reputation across major receivers
- Alert on issues before customer notices
- Investigate and remediate deliverability problems
- Manage IP warmup transitions
- Handle blocklist delisting requests
- Educate customers on best practices
- Coordinate with major receivers when needed
This work is real ongoing operations. It does not stop. New customer onboarding produces new deliverability work. Existing customers produce incident response. Receiver-side changes produce remediation work.
The realistic deliverability operations load: 30-50% of total ESP operational time. New ESPs often underestimate this and end up with deliverability operations consuming founders rather than scaling activities.
Infrastructure operations
Infrastructure does not run itself. Servers need patching. Databases need optimization. Network needs monitoring. Backups need verification. Security needs maintenance. The work scales with infrastructure size, not just with customer count.
New ESPs sometimes assume cloud infrastructure (AWS, Hetzner, etc.) handles this. The cloud handles some of it. The application-layer, database-layer, and ESP-specific infrastructure all still need ongoing operations.
Realistic infrastructure operations load: 10-20% of total operational time at moderate scale. The work is interruption-heavy (incidents trigger investigation, patches require coordination).
Compliance and abuse
ESPs handle email at scale. The scale produces compliance and abuse considerations:
GDPR compliance for EU customers and their EU recipients.
CAN-SPAM compliance for US-facing operations.
CASL for Canadian operations.
Bulk sender requirements (Gmail, Yahoo, Microsoft) and ongoing changes to those requirements.
Abuse handling: spam complaints, blacklist remediation, content policy enforcement.
Subpoena and legal request handling: the ESP becomes a third-party data custodian and must have procedures for responding to legal requests.
Customer offboarding: when customers leave or are terminated, the data handling needs proper procedures.
This work is not glamorous but unavoidable. ESPs that skip compliance and abuse handling face regulatory and reputation consequences.
Realistic compliance and abuse load: 5-10% of total operational time, with spikes when specific regulatory changes happen.
Marketing and customer acquisition
Customers do not appear automatically. The ESP needs marketing and customer acquisition activities ongoing.
Content marketing: writing blog posts, tutorials, comparison articles.
SEO: optimizing for search terms relevant to your customer base.
Partnership development: integrations with other tools your customers use.
Sales: responding to inbound inquiries, demos, qualification, closing.
Account management: keeping existing customers happy, identifying expansion opportunities.
Realistic marketing and acquisition load: 20-30% of total operational time, more if growth is the priority.
The total
Adding up the operational work:
- Customer support: 20-30%
- Deliverability operations: 30-40%
- Infrastructure operations: 10-15%
- Compliance and abuse: 5-10%
- Marketing and acquisition: 20-25%
This is more than 100% of one person’s time. The implication: a single founder cannot run an ESP beyond a small scale. The team needs at least 2-3 people by month 6 to handle everything well, or the operation will be permanently below quality standards.
The economic model
For ESPs that successfully navigate the operational curve, the economics work as follows.
Revenue per customer
Average revenue per customer varies enormously based on positioning:
- Low-touch self-service: €30-€100/month per customer
- Mid-market managed: €200-€800/month per customer
- Enterprise focus: €2,000-€10,000+/month per customer
Total revenue scales with customer count and revenue per customer. A 100-customer ESP at €300 ARPU is €30K monthly recurring revenue. At €1,500 ARPU is €150K MRR.
Cost per customer
Infrastructure cost per customer scales sublinearly. The first customer costs more (fixed infrastructure overhead). Subsequent customers cost less per customer.
Typical infrastructure cost per customer:
- 10 customers: €100/customer
- 50 customers: €50/customer
- 200 customers: €30/customer
- 500+ customers: €20/customer
The sublinear scaling produces improving gross margins as the ESP grows.
Operational cost per customer is harder to scale. Customer support scales linearly with customer count and complexity. Deliverability operations scale with customer count and sending volume.
Typical operational cost per customer (at scale): €40-€80/customer.
Gross margin
Combining: a mid-market ESP at €300 ARPU with €30-€50 infrastructure cost + €40-€80 operational cost = €70-€130 total cost = €170-€230 gross margin = 56-77% gross margin.
The gross margin is reasonable but not exceptional. Compare to mainstream ESPs whose gross margin is reportedly 70-80%. The new ESP can be competitive on price while maintaining good margins, but the absolute revenue is smaller than the mainstream ESP because the customer base is smaller.
Break-even
The break-even point depends on operational cost structure. For an ESP with €15K/month operating cost and €300 ARPU:
- Break-even customers: 50-70 customers depending on cost structure
- Break-even revenue: €15K-€21K MRR
Getting to break-even typically takes 6-12 months for well-executed ESP launches. Slower for less well-positioned launches.
Scale economics
Beyond break-even, the economics improve with scale. The infrastructure cost per customer declines. The operational cost per customer declines (somewhat). The marketing efficiency improves (referrals, content marketing reaches more potential customers).
At 500-1000 customers, the ESP can produce significant cash flow if operated well. €500K-€2M+ in MRR with 60-75% gross margins produces real business value.
When ESP building makes sense
Specific operator profiles align with successful ESP building:
Profile 1: Captive customer base
An operator with existing customers who could be ESP customers has a significant advantage. The customer acquisition cost is low because the customers are already in the relationship. Examples: marketing agencies serving multiple clients, hosting providers with email-sending customers, SaaS platforms with customer email needs.
Profile 2: Vertical specialization
An operator targeting a specific vertical with specific needs that mainstream ESPs do not serve well. Examples: ESP for political campaigns (mainstream ESPs often restrict), ESP for adult industry (mainstream ESPs prohibit), ESP for specific geographic markets (Latin America-focused, MENA-focused, etc.).
Profile 3: Premium services
An operator offering managed services that go beyond what mainstream ESPs provide. Examples: full deliverability management, premium support, custom integration work, regulatory compliance services. Higher ARPU justifies higher cost structure.
Profile 4: Existing infrastructure operator
An operator already running infrastructure (hosting, VPN, etc.) with operational capability and customer relationships. Adding ESP services leverages existing operational capacity.
Profile 5: Long-term technical entrepreneur
An operator with patience for the 6-12 month break-even timeline and capital to fund the journey. The patient capital with technical capability matches the ESP business model.
When ESP building does not make sense
Profiles where ESP building is the wrong choice:
Quick win seekers
Operators looking for fast revenue. ESP building is not a quick win. The 6-12 month timeline does not match impatient capital.
Cost-only competitors
Operators planning to compete primarily on lower pricing than mainstream ESPs. The price war goes to scale, which the new ESP does not have. Differentiation on price alone is a losing position.
Limited technical capability
Operators who cannot operate the infrastructure themselves or oversee operations effectively. ESP operations are technical-intensive. Outsourcing operations to vendors works but eliminates the margin that makes the business interesting.
No customer base or distribution channel
Operators starting completely from scratch with no existing customers and no distribution channel. The customer acquisition cost is too high to overcome at small scale.
Unrealistic financial expectations
Operators expecting hockey-stick growth or expecting to compete with mainstream ESPs in 12 months. The market is mature; the growth is gradual; the patient operators win.
What we provide for ESP builders
We have several specific service offerings for operators building ESPs:
Managed infrastructure
We provide the hosting infrastructure, MTA layer (PowerMTA or Postal), and ongoing infrastructure operations. The ESP builder focuses on platform layer (customer-facing UI, billing, customer management) and customer acquisition.
The pricing: monthly recurring based on volume and customer count. The ESP builder avoids capital investment in infrastructure and engineering time for operations.
IP pool management
We manage the IP allocation, pre-flight validation, warmup operations, and ongoing reputation monitoring. The ESP builder receives IPs that are ready for customer use.
The pricing: per-IP monthly fees plus operational service costs.
Deliverability operations
We provide deliverability monitoring, incident response, blocklist remediation, receiver coordination. The ESP builder focuses on customer relationships while we handle the technical deliverability work.
The pricing: managed service fee per customer or volume-based.
Compliance and abuse handling
We provide procedures and assistance for compliance (GDPR, CAN-SPAM, etc.) and abuse handling. The ESP builder has structured procedures rather than needing to develop them from scratch.
The pricing: included in standard managed services or available separately.
The combination produces an “ESP-in-a-box” offering for operators who want to focus on customer acquisition and business development rather than building from scratch.
The success stories we have seen
Three customers have built successful ESPs on our infrastructure over the past two years. The patterns are informative.
Customer A: marketing agency
A digital marketing agency with 30 existing clients adapted their service to include ESP capabilities. They moved client email infrastructure from various mainstream ESPs to their own platform. Revenue increased because clients pay them more than the previous ESP costs. Margins are good because most clients were already paying for marketing services that included email.
Time to break-even: 4 months (faster than typical because captive customer base).
Current state: 80 clients, €45K MRR, two-person operation with our infrastructure underneath.
Customer B: vertical specialist
A privacy-focused operator targeting customers who could not use mainstream ESPs due to content restrictions. The vertical specialization produced premium pricing (€500-€2,000/month per customer) because alternatives were limited.
Time to break-even: 7 months. Customer acquisition required substantive content marketing and community engagement.
Current state: 35 customers, €60K MRR, three-person operation.
Customer C: regional specialist
An operator targeting Latin American customers. The regional focus produced market position that mainstream ESPs do not occupy strongly. Spanish-language support, regional payment options, and local timezone responsiveness were the differentiators.
Time to break-even: 9 months.
Current state: 120 customers, €38K MRR, two-person operation with regional contractor support.
The customers who pivoted or stopped
Several operators who started ESP builds did not continue. The reasons are also informative.
Pattern 1: Underestimated operations
Operators who started with limited capital and assumed they could handle everything personally. By month 4, the operational load exceeded their capacity. They could not afford to hire help. Customer service quality declined. Customer churn accelerated. The business contracted rather than grew.
These operators typically pivoted to consulting work or rejoined the customer side of ESP relationships.
Pattern 2: Price competition
Operators who positioned primarily on lower pricing than mainstream ESPs. Their pricing did not generate enough margin to fund the operational improvements needed for growth. Customers cared more about quality than price. The business never reached break-even.
Pattern 3: Lost interest
Operators who started with enthusiasm but encountered the realities of the six-month curve and decided ESP building was not what they wanted to do. The operational work is real. Some operators discover they prefer different work patterns and exit the ESP business.
Pattern 4: Acquired by larger player
A few operators built sufficient business that they were acquired by larger ESPs looking for customer base or technical capability. The acquisition provides exit liquidity but ends the independent ESP business.
The honest summary
Building an ESP is viable. The economics work for specific operator profiles. The six-month operational curve is the obstacle most new ESP builders underestimate.
For operators considering this path:
Match your profile to a successful profile pattern. Captive customers, vertical specialization, premium services, existing infrastructure, or patient capital. Without one of these, the path is harder.
Plan capital for the six-month operational curve. Undercapitalization is the leading cause of ESP failure. Better to start with adequate capital than to scramble in month 4 when you realize you need to hire.
Treat operations as a real specialty. Deliverability operations, customer support, and infrastructure operations are not afterthoughts. They are the business. Skipping any of them produces unsustainable operations.
Find partners for what you cannot do yourself. Managed infrastructure providers, deliverability specialists, compliance advisors. The ESP business is more about coordination than building from scratch.
Be realistic about the timeline. Break-even in 6-12 months is realistic for well-executed launches. Faster timelines are exceptional rather than typical.
For operators who match the profile and accept the timeline, building an ESP can produce a substantial business. For operators who do not match the profile or do not accept the timeline, the path is more difficult than it appears from the outside.
We work with operators on both sides of this question. Some come to us building an ESP and we provide the infrastructure. Some come to us considering the ESP path and we help them think through whether it is right for their situation. Both conversations are worth having before committing to the ESP build.
The mainstream ESP market is not as locked-up as it appears. Specific niches are addressable. Specific operator profiles can build sustainable businesses. The path is real. The work is real. The six-month curve is real. Operators who understand all three before committing produce better outcomes than operators who learn them in month 5.