Self-Hosting: The Decision Framework for Builders Who Want Their Data Back
The 8-question decision framework for whether to self-host. Time, threat model, family situation, regulatory needs, and the persona-mapped cluster picks.
The 8-question decision framework for whether self-hosting is right for you, with persona-mapped picks for the right starting stack and the explicit “do not self-host” criteria.
- Self-host if: You have 30+ minutes / week, you want data sovereignty, you have backup discipline, you run 3+ services that would otherwise be SaaS
- Skip self-hosting if: You have under 30 min / week available, no backup discipline, only need 1-2 services, or your team includes daily non-technical users
- Hybrid is fine: Self-host the high-impact services (Vaultwarden, n8n, Plausible). Stay on SaaS for the harder ones (email, real-time team docs)
- Time cost reality: ~30 min / week for the 5-stack starter kit. ~2 hours / week for a 12-service homelab.
- The current starting move: Vaultwarden + Plausible + Uptime Kuma on a Hetzner CCX13, €0 marginal cost beyond the existing instance, near-zero maintenance
The “should I self-host” question gets answered too lazily online. The answers are usually “yes if you are technical” or “no it is too much work” without any structure. We have run self-hosted infrastructure for years and helped a dozen indie founders make this decision. Here is the explicit 8-question framework that we use, with the persona-mapped starting stack and the explicit “do not self-host” criteria.
01At a glance: what we tested
| Question | Self-host signal | Stay on SaaS signal |
|---|---|---|
| Time available per week | 30+ minutes | Under 30 minutes |
| Threat model | Includes “do not trust SaaS provider” | SaaS provider trust is acceptable |
| Backup discipline | You have it or you can build it | You do not and will not |
| Regulatory needs | GDPR / data residency required | No specific compliance needs |
| Existing homelab or VPS | Yes (€0 marginal) | No (must add infra cost) |
| Skill curve interest | Yes (you want to learn) | No (you just want it done) |
| Failure tolerance | Comfortable troubleshooting | Need vendor SLA |
| Team size and shape | 1-3 person team | 10+ person team with non-technical members |
02The 8 questions, in order
Score yourself across the 8 questions. 6+ self-host signals: clear yes. 4-5: hybrid. 3 or fewer: stay on SaaS.
Buy if: not applicable. Skip if: not applicable.
Question 1 (Time): can you commit 30 minutes per week to maintenance? If no, stop here. Question 2 (Threat model): does your threat model specifically exclude trusting a SaaS provider? Question 3 (Backup discipline): do you have or can you build a tested backup pipeline (3-2-1 rule, restore tested at least quarterly)? Question 4 (Regulatory): does GDPR, HIPAA-equivalent, or data-residency law require self-hosting? Question 5 (Existing infra): do you already run a Hetzner instance or a mini-PC? Question 6 (Skill curve): are you genuinely interested in learning the stack? Question 7 (Failure tolerance): are you comfortable troubleshooting at 11 PM when something breaks? Question 8 (Team shape): is your team 1-3 technical people or 10+ with non-technical daily users? Score yourself: 6+ self-host signals = clear yes. 4-5 = hybrid. 3 or fewer = stay on SaaS.
03Persona-mapped starting stacks
Four personas with the right starting stack for each. Most builders fit one of these.
Buy if: not applicable. Skip if: not applicable.
Persona 1 (solo indie founder, pre-revenue): Vaultwarden + Plausible + Uptime Kuma. Hetzner CCX13. Total time: 2 hours setup, 5 minutes / week maintenance. Read the starter-kit guide. Persona 2 (early-MRR indie founder, 1-3 team): add n8n + Linkwarden. 8 hours setup total, 30 minutes / week maintenance. Persona 3 (post-PMF, 5-10 team): add Nextcloud + Plex / Jellyfin. Step up to mini-PC at home or Hetzner CCX23. 4-8 hours / month maintenance. Persona 4 (digital nomad / lives + travels): cloud-only on Hetzner. Tailscale + Cloudflare Tunnel for remote access. Skip Plex / Jellyfin (4K transcoding tight on cloud). Add Home Assistant if you have a home base with smart devices. The right pick depends on which persona you are, not on general “should I self-host” reasoning.
04When NOT to self-host (the explicit criteria)
Three scenarios where self-hosting is the wrong choice regardless of how technical you are.
Buy if: not applicable. Skip if: not applicable.
Scenario 1: you have under 30 minutes / week for maintenance. Even the lightest self-hosting setup needs occasional update checks, backup verification, and troubleshooting. Below 30 minutes / week, you will skip these and the eventual incident will cost more than the SaaS subscription would have. Scenario 2: your team includes daily non-technical users (5+ people who interact with the service every day and cannot troubleshoot). UX polish gap on self-hosted alternatives means daily friction; the SaaS premium pays back in team productivity. Scenario 3: no backup discipline. A self-hosted password manager without backups is one disk-failure from losing all your credentials. A self-hosted Nextcloud without backups is one ransomware-attack from losing all your business documents. If you will not commit to backup discipline, do not self-host the high-stakes services.
05Which option should you pick?
Pick by your situation
- Can commit 30 min / week and have backup discipline? → Self-host (start with Vaultwarden + Plausible + Uptime Kuma)
- Solo founder, pre-revenue, just want to save money? → Hybrid: self-host Vaultwarden, stay on Bitwarden Premium for the rest
- Team has daily non-technical users? → Stay on SaaS for shared services; self-host personal-only services
- Live + travel, want everything online? → Cloud-only Hetzner CCX13 + Tailscale + Cloudflare Tunnel
- GDPR or data-residency requirement? → Self-host on EU-region Hetzner; do not skip this
- No backup discipline and will not build it? → Stay on SaaS. Self-hosting will lose your data
06FAQ
What is the lowest-stakes way to start self-hosting?
Vaultwarden + Plausible + Uptime Kuma on a Hetzner CCX13. Total marginal cost: €0 if you already run Hetzner; €5 / month if not. Total maintenance: 5 minutes / week. Highest ROI / lowest risk combination.
How long does it take to feel competent self-hosting?
About 3 months of running 3-5 services. The first month is rough. You hit unexpected issues and learn the operational patterns. Months 2-3, things settle. After 3 months, adding a new service feels routine.
What if my homelab catches fire?
Off-site backup is the answer. We back up to Backblaze B2 ($5 / TB / month, encrypted with our own key). If our homelab burns down tomorrow, we restore to a fresh Hetzner instance from Backblaze in under 4 hours. Test the restore once per quarter; it is the most important self-hosting habit.
Can I self-host without a homelab. Just on cloud?
Yes. Most of the indie-founder stack runs on a Hetzner CCX13 (€4.85 / month). Plex / Jellyfin and Immich benefit from local hardware (transcoding, storage cost), but everything else is fine on cloud. We have run cloud-only setups for 18+ months without issue.
What happens if I get hit by a bus?
Critical question that most self-hosting guides skip. Set up a Vaultwarden Emergency Access contact who can recover your accounts. Document your stack in a single Markdown file your partner or executor can read. Use a fire-safe for the master password and recovery keys. Self-hosting is not about being the only person who can access your data; it is about choosing who else can.
07WikiWalls verdict
WikiWalls verdict. Self-host if you have time, threat model, backup discipline, and either an existing instance or a willingness to run one. Skip if any of those is missing. The hybrid path (self-host high-impact services, stay on SaaS for the rest) is underrated and right for most indie founders.
Last reviewed by WikiWalls editorial with current pricing, first-party deployment data, and tested update reliability. Recommendations are editorially independent.
Last reviewed by WikiWalls editorial. Recommendations are editorially independent. Methodology: /test-methodology/. Editorial standards: /editorial-standards/.