What SaaS Maintenance Actually Costs (Real Numbers)

The biggest surprise for most founders happens six months after launch: the SaaS they built for $60,000 now costs $8,000 to $15,000 a month to keep alive. Nobody talked about that number in the initial planning conversation because it is not a satisfying line item to include when you are pitching investors on speed. This piece breaks down what SaaS maintenance actually costs in real numbers based on the fifty-plus products I have watched through their first two years, including several that shipped inside a QwiklyLaunch 45-day scope. I will cover infrastructure, engineering, support, security, and the hidden line items nobody warns you about. Numbers are for a typical B2B SaaS with 500 to 5,000 users, single-region deployment, and a small team of two to five engineers. Adjust up or down for your scale, but the categories are the same at every stage.

Infrastructure: The Base Layer

Infrastructure is the most visible cost and usually the smallest one. A single-region SaaS with a Postgres database, a Redis cache, and a Node or Python API on a container platform runs $200 to $800 a month at 500 users. At 5,000 users you are looking at $1,500 to $4,000 a month depending on how chatty your app is and whether you have real-time features like websockets.

The typical breakdown at 5,000 users is: application hosting on Vercel, Render, or Fly at $300 to $800, Postgres on Supabase or Neon at $200 to $600, Redis and background jobs at $100 to $300, object storage on S3 or R2 at $50 to $200, and CDN plus egress at $100 to $500. Add another $200 to $500 for observability tools like Sentry and Datadog if you use paid tiers.

The Egress Trap

Egress bandwidth is the line item that surprises founders when they hit product-market fit. AWS charges $0.09 per gigabyte outbound, which is fine at low volume and terrifying at scale. A product that streams video, serves large images, or exports big reports can rack up $2,000 a month in egress alone. Use Cloudflare R2 or Bunny CDN if egress is significant in your workload, they are cheaper by an order of magnitude.

Engineering: The Biggest Line Item

Engineering time is the largest cost by a wide margin. A single mid-level engineer to keep a SaaS running, ship small features, and respond to on-call runs $8,000 to $18,000 a month depending on region and full-time versus contract. A senior engineer runs $15,000 to $30,000 a month. Most founders underestimate this because they think of engineering as feature work, not maintenance.

Maintenance-only engineering, without new features, is typically twenty to forty percent of a full-time engineer. That means one senior engineer split across maintenance and feature work is the minimum viable staffing for a live SaaS. If you have no dedicated maintenance capacity, feature work will slow to a crawl within six months as bugs and technical debt pile up.

On-Call And Incident Response

On-call is a hidden engineering cost. Rotating one engineer on primary and one on secondary through a 24-hour cycle costs roughly ten to fifteen percent of their total capacity, even in weeks with no incidents. Add another five percent for post-incident writeups, retros, and follow-up fixes. Budget on-call as a real line item, not as an afterthought, and consider outsourcing it to a firm like our devops and cloud practice if the team is too small to sustain a rotation.

Support: More Than A Shared Inbox

Customer support scales with users, not with revenue. At 500 users you can usually get by with the founder answering emails for an hour a day. At 2,000 users you need a part-time support person at $2,000 to $4,000 a month. At 5,000 users you need a full-time support lead at $4,000 to $8,000 a month plus tooling.

Support tooling adds another $200 to $600 a month: Intercom or Front for the inbox, a help center like Helpjuice or a hosted docs setup, and a status page like Instatus or Statuspage. If you offer live chat, expect a much higher volume of low-value tickets and staff accordingly. Async support through email or in-app messaging scales better and is cheaper per ticket.

Ticket Volume Benchmarks

Expect roughly two to five tickets per hundred active users per week for a mature SaaS with good docs and onboarding. Higher than that and your product has UX debt that support cannot fix. Track ticket volume by user cohort and by feature area, and feed the top three ticket categories back into engineering as the next quarter's UX priorities.

Security: The Cost Of Not Getting Breached

Security is not a feature, it is ongoing work. Baseline security costs include a bug bounty program at $200 to $1,000 a month, dependency scanning through Snyk or GitHub Advanced Security at $100 to $500, secrets management through Doppler or AWS Secrets Manager at $50 to $200, and a security review pass every six months at $3,000 to $10,000 per pass.

Add SOC 2 costs if you sell to enterprise. Initial audit runs $15,000 to $40,000 and ongoing compliance software like Vanta or Drata runs $500 to $2,000 a month. SOC 2 is not maintenance, it is a sales prerequisite, and if you sell upmarket you need to budget it as a fixed cost of doing business.

The Cost Of Being Breached

The direct cost of a data breach for a small SaaS averages $150,000 to $500,000 when you include incident response, legal, customer notification, and churn. That number does not include reputation damage or lost future revenue. Preventive security spending of $2,000 to $5,000 a month is cheap by comparison, and skipping it is one of the fastest ways to kill a SaaS.

Third-Party Services And Add-Ons

Third-party services accumulate quietly. A typical B2B SaaS at 2,000 users pays: Stripe fees at 2.9 percent plus $0.30 per transaction, email through Postmark or Resend at $50 to $300, transactional SMS through Twilio at $100 to $500, analytics through PostHog or Mixpanel at $100 to $800, error tracking through Sentry at $100 to $400, and feature flags through LaunchDarkly or Statsig at $100 to $500.

These add up to $500 to $2,500 a month easily, and they usually grow faster than infrastructure because they are priced per user or per event. Audit your third-party stack every quarter and cancel or downgrade anything with under twenty percent utilization. This audit typically saves fifteen to thirty percent of the third-party bill.

Dependency Upgrades And Technical Debt

Dependency upgrades are the maintenance work nobody schedules but everybody needs. Node, Python, Postgres, and framework upgrades typically eat five to fifteen percent of engineering capacity across a year. Skipping upgrades makes future upgrades exponentially more expensive because breaking changes compound.

Budget a quarterly dependency upgrade sprint of one to two weeks of engineering time. This keeps you within one major version of the ecosystem and prevents the doomsday scenario where you cannot ship because a critical library dropped support for your version. Our API and backend development playbook covers structured upgrade cadences in detail.

The Real Monthly Number

For a typical B2B SaaS at 500 to 2,000 users with a small team, expect $8,000 to $18,000 a month in all-in maintenance costs. At 2,000 to 5,000 users the number moves to $15,000 to $35,000 a month. This includes one to two engineers, part-time support, infrastructure, security, and third-party services.

Founders who plan for these numbers upfront do not get surprised. Founders who do not plan get surprised at month six, when the burn rate is double what they expected and revenue has not caught up. Budget maintenance as a fixed cost from day one, not as a variable that scales with revenue, because it does not scale that way. It scales with users and with time.

What QwiklyLaunch Includes And Excludes

The QwiklyLaunch 45-day scope covers the build and the launch. It does not include ongoing maintenance because maintenance is a different kind of engagement with different economics. After launch, we offer a maintenance retainer that covers bug fixes, dependency upgrades, security patching, and small feature work at a predictable monthly rate. Most founders start with a three-month post-launch retainer and then decide whether to hire in-house or extend.

If you want a maintenance quote or a discussion about what your SaaS will actually cost to run at your projected scale, reach out through the contact page. You can also see examples of live SaaS products we currently maintain in our projects gallery.