Self-Hosted vs SaaS: The True Cost of Per-Seat Software

Short answer: per-seat SaaS is cheaper to start, but self-hosted open-source software usually wins on total cost once you have more than ~15–30 users — because you pay for infrastructure, not headcount. This guide shows exactly how the two cost curves differ, where they cross, the hidden costs of each, and how to decide for a specific tool.

SaaS pricing feels painless at first — a few dollars per user, billed monthly, no setup. The pain arrives quietly, as your team grows and your software bill grows with it. Understanding the two cost curves is the key to knowing when ownership beats renting.

How SaaS pricing actually scales

Per-seat SaaS charges you for every user, every month, forever. A $40/user tool costs $480/year for one person — and $48,000/year once 100 people use it. Your software bill grows with your team even when each person's usage is identical. That's the core problem: your cost scales with success, not with value.

There are second-order costs, too. "Premium" tiers gate the features you actually need behind higher per-seat prices. Add-ons (extra storage, API access, SSO — the infamous "SSO tax") pile on. And because pricing is per seat, you start rationing logins, which undermines the tool's whole point.

What self-hosting costs instead

Self-hosted open-source software has a different shape: a one-time setup cost, then ongoing infrastructure (a server on AWS or your own cloud) measured in dollars per month, not dollars per user. A platform serving 100 people often runs on the same modest server as one serving 10.

Cost driver	Per-seat SaaS	Self-hosted open source
Licensing	$/user/month (grows with team)	Free
Infrastructure	Included	You pay (often $20–$200/mo)
Setup	Low	One-time, higher
Customization	Limited / paid tiers	Unlimited
Data ownership	Vendor	You
SSO / API	Often an upsell	Included

The two cost curves, side by side

Picture two lines on a graph over three years. SaaS is a line sloping steadily upward with every new hire. Self-hosting is a step (the setup cost) followed by a nearly flat line (hosting). Early on, SaaS is below — it's cheaper to start. At some point the rising SaaS line crosses the flat self-hosted line, and after that the gap widens every month.

The crossover point

For most tools, self-hosting becomes cheaper somewhere between 15 and 30 active users, and the gap widens fast after that. A simpler rule of thumb: if you're spending $1,000–$3,000/month on a single SaaS tool, it's worth pricing the self-hosted alternative.

A worked example. A 40-person team on a $75/user CRM pays $36,000/year, rising with every hire. A self-hosted, customized open-source CRM might cost $20k to set up and tailor, plus ~$150/month to host. Year one: roughly break-even. Year two: you save ~$34,000. Year three: the same — and you own it. Over three years that's tens of thousands of dollars kept, plus a system shaped to your process.

The hidden costs (on both sides)

Be honest about both columns.

Self-hosting's real costs:

• Setup — getting it deployed, configured, and integrated.
• Hosting — the monthly server/database/storage bill.
• Updates — keeping the app and dependencies patched.
• Backups & monitoring — automated, tested, and alerted.

The important point: a partner can turn all of these into a fixed, predictable monthly cost — you don't have to staff a platform team. (See self-hosting on AWS.)

SaaS's hidden costs:

• Annual price increases you can't control.
• Feature gating that forces tier upgrades.
• Per-seat rationing that limits adoption.
• Exit cost — the time and risk of migrating off if you ever leave.

Beyond price: what you also gain by owning

Cost is the headline, but ownership brings more:

• No lock-in. Open-source cores mean you can change or move providers anytime.
• Customization. Shape the tool to your workflow instead of the reverse.
• Data ownership. Everything stays in infrastructure you control — meaningful for compliance-sensitive industries.
• No per-seat anxiety. Give every team member access without watching the meter.

A category-by-category cost reality

The per-seat problem is worst in the categories where you have the most logins. A rough sense of where self-hosting tends to pay off fastest:

Category	Per-seat SaaS	Self-hosted alternative	Pays off fastest when…
CRM	Salesforce, HubSpot	Twenty	Sales/CS team grows past ~15
Support	Intercom, Zendesk	Chatwoot	Many agents or high ticket volume
Analytics	Amplitude, Mixpanel	PostHog	Usage-based bills spike
Scheduling	Calendly	Cal.com	Whole org needs booking links
Project mgmt	Jira	Plane	Large team, many viewers

The pattern is consistent: the more people who need access, the stronger the case for owning it. See the full open-source alternatives guide.

Is self-hosting riskier? (Security & reliability)

A fair concern — and the honest answer is "not when it's set up properly." Self-hosting on a major cloud like AWS gives you the same security primitives the big SaaS vendors use: network isolation, encryption, managed databases, and automated backups. Reliability comes from sensible architecture (a right-sized server, a managed database, monitoring) rather than heroics.

What you trade is convenience for control. SaaS handles ops invisibly; self-hosting puts ops in your hands — or your partner's. The mitigation is simple: have someone own updates, backups, and monitoring. Done once, properly, a self-hosted platform is as dependable as the SaaS it replaced, and your data never leaves accounts you control. (More in self-hosting on AWS.)

When SaaS still wins

Managed SaaS is the right call when:

• Your team is small and per-seat cost is trivial.
• The need is commodity and a mature product already fits.
• You genuinely don't want to think about hosting, even managed.
• The tool is not core to how you operate or compete.

The honest answer is that it depends on scale and how much the tool touches your differentiated work. Don't self-host on principle — do it where the math and the strategy agree.

Common objections to self-hosting, answered

"We don't have engineers to run it." You don't need to. A partner deploys it, automates updates and backups, monitors it, and hands over documentation — turning ops into a predictable monthly cost. Many teams self-host without a single in-house DevOps hire.

"Isn't it less reliable than SaaS?" Not when it's built on a major cloud with sensible architecture and monitoring. The same infrastructure that powers large SaaS products is available to you on AWS.

"What about updates and new features?" Mature open-source projects ship updates regularly; your partner applies them. You get a steady stream of improvements without a rising bill.

"It sounds like more work upfront." It is — that's the trade. You pay a one-time setup cost in exchange for years of lower, flatter costs and full ownership. The three-year math is what makes it worthwhile.

What "self-hosted" does not mean

A few clarifications, because the term scares people more than it should:

• It does not mean a server in your closet. It means your cloud account (AWS, GCP, Azure) — managed infrastructure, just owned by you.
• It does not mean building software from scratch. You're running mature, proven open-source applications.
• It does not mean you're on your own. A partner can own the operational side entirely.
• It does not mean less security. Often the opposite, since your data stays in accounts you control.

Self-hosting in 2026 is closer to "managed hosting of software you own" than the DIY image the word conjures.

A 4-step way to decide for one tool

1. Total the annual per-seat cost (users × price × 12), and project it across three years with expected hiring.
2. Estimate the self-hosted cost — setup + 36 months of hosting/maintenance.
3. Compare the three-year totals, not the monthly headline.
4. Weigh the non-cost factors — customization, data ownership, lock-in — and decide.

A 3-year total-cost worksheet

Numbers beat opinions. Here's a simple worksheet to compare a specific tool honestly over three years — the timeframe where the curves diverge.

SaaS column:

1. Current users × monthly price × 12 = year-one cost.
2. Add your expected hiring — recalculate years two and three at the larger headcount.
3. Add known price increases (assume ~5–10%/year if unsure) and any add-on/tier costs.
4. Sum all three years = SaaS three-year total.

Self-hosted column:

1. One-time setup/customization estimate.
2. Monthly hosting × 36 (often $20–$200/mo depending on scale).
3. Maintenance (a small retainer or in-house time) × 36.
4. Sum = self-hosted three-year total.

Then compare the two totals — and weigh the non-cost factors below. For a mid-sized team on a high-seat tool, the self-hosted total is frequently half the SaaS total over three years, with ownership on top.

How a switch actually happens

Worried the migration will be painful? A typical, low-risk switch looks like this:

1. Week 0 — Scope & price. Confirm the open-source target fits and agree a fixed setup cost.
2. Weeks 1–2 — Deploy & configure on your cloud, with backups and monitoring.
3. Weeks 2–3 — Customize & migrate data. Tailor it to your workflow and import your records (see the data migration playbook).
4. Week 3–4 — Parallel run. Keep both systems live briefly to validate.
5. Cutover. Switch during a quiet window; archive the old export.

Most switches for a single tool wrap in 3–6 weeks, and you're left with a system you own and a bill that stops growing with your team.

The bottom line

Per-seat SaaS is a rational way to start — low commitment, no setup. But it's a renting model, and rent never stops or shrinks; it grows with every hire. Self-hosting open source is a buying model: a higher upfront cost, then a flat, modest bill and an asset you own. The crossover usually arrives faster than people expect, especially for high-seat tools like CRM and support. The disciplined move is to keep renting the cheap, commodity tools and buy the expensive ones you'll use for years.

Key takeaways

• Per-seat SaaS cost scales with your team; self-hosting costs scale with infrastructure, not users.
• The crossover where self-hosting wins is often 15–30 users, or ~$1,000–$3,000/month in SaaS spend.
• Self-hosting's real costs (setup, hosting, updates, backups) can become a fixed monthly cost with a partner.
• Ownership also brings no lock-in, full customization, and data control.
• Keep SaaS for small teams and commodity needs; switch where the three-year math favors ownership.

If you want a real number for your situation, tell us what you're paying and we'll price the self-hosted equivalent — set up and maintained on your cloud.