Your finance team scheduled a budget review. They pulled up the software spend dashboard, and there it was: collaboration tools that started at $500/month last year now cost $3,200. Nobody added features. Headcount increased from 40 to 65. The SaaS calculator promised “predictable pricing.” It didn’t mention the part where crossing user thresholds triggers tier jumps that double your monthly spend.
This scenario repeats across thousands of companies every quarter. The initial pricing looks reasonable. “Pay as you grow, scale seamlessly, no infrastructure headaches.” Then growth happens—the kind where your product succeeds, and your team expands—and software costs scale faster than revenue. Engineering starts asking questions. Finance wants alternatives. Someone mentions self-hosting, and the vendor sales rep warns about “hidden infrastructure costs.”
Let’s examine what these costs actually total over time, using research from Forrester and Gartner on total cost of ownership (TCO), alongside real pricing data from AWS and SaaS vendors.
How SaaS Pricing Actually Works
SaaS vendors design pricing calculators to show best-case scenarios. Multiply the monthly subscription fee by the expected number of users to get your total cost. Simple math. But that calculator doesn’t account for the compounding effects of per-seat pricing, data export fees when you need to migrate, feature paywalls that force enterprise-tier upgrades, or the fact that “active user” definitions change at contract renewal.
According to Bain & Company’s analysis of per-seat pricing, while the per-user model remains dominant (57% of SaaS companies use it as their primary model), its dominance is declining. The model creates predictable cost escalation: as your team grows, costs scale linearly or worse, often hitting tier thresholds that create step-function increases in your monthly bill.
The psychology behind this is deliberate. Low entry pricing hooks your team. You build workflows around the tool. You integrate it with other systems. Switching costs accumulate—not just monetary costs, but workflow disruption, retraining time, and migration risk. By the time you recognize the total cost trajectory, you’re locked in. The pricing team designed for this outcome.
Compare this to infrastructure costs, which vendors position as “unpredictable” and “requiring specialized expertise.” Reality check: A server that supports 10 users typically scales to 50 users without incurring proportional scaling costs. Infrastructure scales logarithmically—costs grow slower than usage increases—while SaaS pricing scales linearly with tier jumps that create sudden cost spikes. Based on AWS Pricing Calculator data, a $50/month server instance for 10 users might become a $100/month instance for 50 users. Meanwhile, that SaaS collaboration tool went from $80 to $1,000 monthly.
What Gartner Research Shows About TCO
Gartner’s TCO research reveals that personnel costs for maintaining on-premises systems account for 50-85% of total application costs. This finding gets used by SaaS vendors to justify their pricing, but it’s worth examining the context. Gartner estimates that over 75% of IT budgets go toward maintaining existing systems—but that includes legacy on-premise installations that predate modern DevOps practices, containerization, and infrastructure-as-code tools.
Modern self-hosted open source applications don’t require dedicated system administrators per application. Using Docker Compose or Kubernetes, a single operations team member can manage dozens of containerized applications. Stack Overflow’s 2024 Developer Survey shows that over 65,000 developers reported increased adoption of containerization and cloud-native practices, making self-hosting significantly less labor-intensive than traditional on-premise deployments.
The Forrester TCO comparison underscores that software vendors need to address customer concerns about SaaS out-of-pocket costs by presenting a transparent total cost of ownership. The research notes that developers using cloud-based collaboration tools saved 1.5 hours per day compared to on-premise versions—but that efficiency gain came from better software design, not from the SaaS delivery model itself. Open source alternatives like Mattermost deliver the same productivity improvements through modern architecture, without the per-seat pricing trap.
Real Infrastructure Costs
Let’s run actual numbers using AWS infrastructure pricing. For a team of 50 people running core business applications:
Collaboration Platform (Mattermost):
- AWS EC2 instance (t3.large): $60/month
- RDS PostgreSQL (db.t3.medium): $70/month
- S3 storage (500GB): $12/month
- Data transfer: $15/month
- Monthly total: $157
Content Management (Strapi or Payload CMS):
- AWS EC2 instance (t3.medium): $30/month
- RDS PostgreSQL (db.t3.small): $35/month
- S3 storage (1TB): $24/month
- CloudFront CDN: $20/month
- Monthly total: $109
Project Management (Plane):
- AWS EC2 instance (t3.medium): $30/month
- RDS PostgreSQL (db.t3.small): $35/month
- S3 storage (100GB): $2/month
- Monthly total: $67
Design Collaboration (Penpot):
- AWS EC2 instance (t3.medium): $30/month
- RDS PostgreSQL (db.t3.small): $35/month
- S3 storage (500GB): $12/month
- Monthly total: $77
Combined infrastructure: $410/month for four core business applications serving 50 users.
Now compare SaaS equivalents at 50 users:
- Slack (Business+): $12.50/user = $625/month
- Contentful (Team): $489/month base + API overages
- Linear (Business): $19/user = $950/month
- Figma (Professional): $15/user = $750/month
Combined SaaS cost: $2,814/month minimum
The gap widens over time:
| Timeline | Self-Hosted (AWS) | SaaS Equivalent | Difference |
|---|---|---|---|
| Year 1 | $4,920 | $33,768 | -$28,848 |
| Year 3 | $14,760 | $101,304 | -$86,544 |
| Year 5 | $24,600 | $168,840 | -$144,240 |
These calculations assume no SaaS price increases (unrealistic—CIO Magazine reports that SaaS vendors are raising prices 9-25% annually while IT budgets grow at 2.8%) and no team growth beyond 50 people (also unrealistic for successful companies).
The Personnel Cost Question
SaaS vendors emphasize personnel costs: “You’ll need engineers to maintain self-hosted systems.” Let’s examine this honestly. Modern containerized applications require minimal ongoing maintenance. Updates happen via CI/CD pipelines. Monitoring uses standard tools (Prometheus, Grafana). Backups run automatically.
For a company with 50 people, you likely already have a DevOps engineer or platform team managing your cloud infrastructure, CI/CD, and production systems. Adding four containerized applications to that infrastructure represents incremental work, not full-time headcount. If we generously estimate 5 hours per application per month for updates, monitoring, and maintenance, that’s 20 hours per month—half a week of engineering time.
At a $150,000 annual salary (senior DevOps engineer), that’s roughly $1,400/month in labor costs. Add this to the $410 infrastructure cost: $1,810/month total. Still 36% cheaper than the $2,814 SaaS baseline, and this gap grows as SaaS pricing scales with headcount while infrastructure and labor costs remain relatively flat.
What This Means for Your Business
The TCO analysis reveals several patterns:
Year 1: SaaS appears more expensive but arguably worth it for speed and simplicity while establishing product-market fit. The $28,848 premium buys focus on core product development rather than infrastructure setup.
Years 2-3: The cost differential becomes harder to justify. Your engineering team has gained operational maturity. The applications you need are well-understood. The $86,544 difference over three years could fund a platform engineering hire entirely.
Years 4-5: SaaS costs compound while infrastructure costs plateau. The $144,240 difference represents real money—enough for two additional engineers, expanded cloud infrastructure, or reinvestment in product development.
The break-even calculation depends on your team’s capabilities and priorities. If you have strong DevOps practices, containerization experience, and modern infrastructure automation, the TCO heavily favors self-hosting. If you’re a five-person startup racing to validate product-market fit, SaaS makes sense despite higher costs.
Making the Decision
Use this framework for evaluating SaaS versus self-hosted options:
Consider SaaS when:
- Your team lacks DevOps/infrastructure expertise
- You’re pre-product-market fit and need maximum development velocity
- The application has complex compliance requirements better handled by vendors
- Vendor-specific features provide a genuine competitive advantage
Consider self-hosting when:
- You have platform engineering capability or plan to build it
- Your team has grown beyond 30-50 people (per-seat pricing becomes punitive)
- You’re experiencing SaaS cost growth outpacing budget growth
- Data sovereignty or security requirements favor self-managed infrastructure
- The open source alternative provides 80%+ of the features you actually use
The Stack Overflow 2024 survey data shows growing developer comfort with containerization and cloud-native architecture. The tools for self-hosting have matured significantly. Docker Compose deployments take minutes. Kubernetes operators handle complex application lifecycles automatically. Infrastructure-as-code eliminates configuration drift.
Moving Forward
SaaS pricing calculators show you the entry price, not the five-year total. Per-seat models create cost escalation that outpaces business growth. The “no maintenance required” pitch obscures the reality that modern DevOps practices make self-hosting far less burdensome than legacy on-premise installations.
Run your own TCO analysis using actual pricing from AWS Calculator, your current SaaS invoices, and realistic estimates of engineering time. Factor in the cost trajectory, not just current spend. Consider what that cost difference enables over five years—additional engineering headcount, expanded infrastructure, product reinvestment, or simply better unit economics.
The total cost of ownership extends beyond monthly invoices. It includes pricing lock-in, vendor dependencies, data export challenges, and the flexibility to optimize infrastructure as your needs evolve. Those factors don’t appear in the calculator, but they determine whether your software costs enable growth or constrain it.