The realistic comparison most orgs face. "We're on Atlassian + Slack + Salesforce + Datadog. What do we get + what do we give up by moving to PMA?"
For a 30-person org doing software / professional services, the typical
cloud bill looks something like:
| Tool | Replaces | List price/user/mo | 30 users |
|---|---|---|---|
| Atlassian Jira | Issue tracking | $7-15 | $210-450 |
| Atlassian Confluence | Docs | $5-10 | $150-300 |
| Slack | Chat | $7-15 | $210-450 |
| Salesforce Essentials | CRM | $25 | $750 |
| Salesforce + Sales Cloud | full CRM/ERP | $80-300 | $2400-9000 |
| Zendesk | Customer support | $19-89 | $570-2670 |
| Mailchimp | Marketing automation | $0-300 flat | $50-300 |
| Datadog | Observability | $15-30 | $450-900 |
| Google Workspace | Email + docs | $6-18 | $180-540 |
Conservative total: $3,000/month = $36,000/year.
Aggressive total (Salesforce full + Datadog full + Zendesk pro): $10,000+/month = $120,000+/year.
| Cloud SaaS | PMA equivalent | Coverage |
|---|---|---|
| Atlassian Jira | Redmine | 90% — issue tracking, sprints, epics, custom workflows. Less polished UI. |
| Atlassian Confluence | Wiki.js | 95% — markdown wiki, page hierarchy, search, page history |
| Slack | Mattermost | 95% — channels, threads, DMs, calls (with Mattermost Calls plugin), bots, integrations |
| Salesforce Essentials | EspoCRM | 80% — contacts, leads, opportunities, basic pipeline |
| Salesforce Sales Cloud / NetSuite | ERPNext | 70-85% — full ERP, accounts, sales, purchasing, stock, HR, projects. Less polish than $100M-vendor product. |
| Zendesk | Zammad | 90% — tickets, email-in, customer organisations, SLAs, KB |
| Mailchimp / HubSpot Marketing | Mautic | 80% — drip campaigns, segments, landing pages, A/B testing |
| Datadog / New Relic | Grafana + Prometheus + Loki | 75% — metrics + logs + alerts; lacks APM trace-correlation polish |
| Google Workspace email | Postal (outbound) + ? | 30% — outbound SMTP only; inbound needs separate solution |
| Google Drive / Notion docs | Wiki.js + AFFiNE (optional) | 60% — collaborative docs less mature than Notion |
| Authentication (Okta / Auth0) | Authentik | 95% — OAuth, OIDC, SAML, MFA, LDAP |
~80-90% feature parity across the stack. Some tools (ERPNext vs
NetSuite) trade polish for control + cost. Others (Mattermost vs Slack) are
genuinely comparable.
For a 30-person org running PMA:
| Cost type | Cloud stack | PMA |
|---|---|---|
| Software licenses | $36,000-120,000/year | $0 |
| Hosting (1 prod + 1 staging host) | $0 (included in SaaS) | $1,200-3,600/year |
| Operator time (setup) | 0-2 weeks of integration work | 4-8 weeks of setup work (one-time) |
| Operator time (ongoing) | ~2 hours/month for admin | ~4-8 hours/month for ops |
| Customisation / API work | Per-product limits + paid tiers | Built into framework |
| Vendor lock-in | High (data export limitations) | None (git + open formats) |
| Year-1 total (rough) | $36-120K | $5-15K |
| Year-N total (rough) | $36-120K (recurring) | $4-12K (recurring) |
The break-even point for PMA is typically 6-12 months after the initial
setup investment. After that, every year is multiple-tens-of-thousands of
dollars saved.
Cost. Obvious. 80-95% reduction in software-license spend.
Data ownership. Your data is in Postgres + MariaDB + named Docker
volumes on a host you control. No "export limitations" or "data retention
limits" or "API call quotas".
No vendor risk. SaaS vendors get acquired, change pricing, deprecate
features, force migrations. Open-source software doesn't.
Customisation depth. Every PMA service is fully open + extensible. You
can write a Redmine plugin or a Wiki.js module or an n8n workflow that does
exactly what you need — not what a SaaS vendor decided to ship.
No tracking / surveillance. Your customer conversations in Zammad,
your strategic plans in Wiki.js, your sales pipeline in EspoCRM — none of
it is being mined by a third-party ML team for "product improvements".
Compliance simplicity. Data residency, data minimisation, deletion
guarantees — you own them all because you own the host.
AI-agent integration. PMA's MCP gateway. Your AI agents can call your
ticketing / chat / CRM with attribution. SaaS APIs sort-of-support this
but rarely with first-class operator-grade auth.
Polish in some products. ERPNext is less polished than NetSuite.
Mattermost's mobile app is less smooth than Slack's. Zammad's UI is more
"functional" than "delightful". Across most tools this gap is small; in
ERPNext vs NetSuite specifically it's noticeable.
Cutting-edge AI features. Notion AI, Slack AI, Atlassian Intelligence
— SaaS vendors ship these regularly. PMA services can integrate AI (via
MCP) but you're wiring it up, not getting it built-in.
Mobile app polish. Most PMA services have OK web UIs that work on
mobile. Native mobile apps where available (Mattermost, n8n) are less
polished than Slack / iOS-grade.
"Just works" first day. SaaS is up in 5 minutes. PMA is up in 30
minutes for a basic stack but production-ready takes weeks of operator
investment.
Vendor support phone number. When something goes wrong at 2 AM,
SaaS vendors have an oncall team. PMA has you + Redmine issue tracker +
community.
Some niche capabilities. Datadog APM auto-instrumentation, Salesforce
Lightning Process Builder, Zendesk's reportable AI Insights, Notion's
real-time co-editing — each is genuinely better in their respective SaaS
than PMA's alternative.
Few orgs go 100% in either direction. Common hybrid patterns:
The point is: PMA is a tool, not an ideology. Pick what makes sense per
domain.
Don't try to switch everything at once. Sequence:
See /pma/cookbook/migrate-from-cloud-saas
for per-migration recipes.
Most orgs can save 50-90% of their software-license spend by moving
~70-80% of their stack to PMA, while keeping 1-2 cloud SaaS tools where
the gap matters (sales CRM, APM, mobile-critical tools).
The break-even is 6-12 months. The annual savings after that are
typically large enough to fund an engineer's salary.
The right question isn't "PMA or SaaS" — it's "which tools am I
overpaying for, and what would it take to bring those in-house?"
/essays/pma-magic — long-form on the cost model + the deeper "why" beyond cost./pma/cookbook/migrate-from-cloud-saas — Jira → Redmine walked end-to-end./pma/learn/00-start-here — entry to the build-it-yourself ladder./pma/cookbook/run-pma-as-a-cooperative + /pma/cookbook/share-pma-with-clients — multi-tenant patterns.