SaaS or Self-Hosting? What I Would Actually Use.

April 22, 2026 · 6 min read

tl;dr

I’ve been developing digital signage software since 2012. Meanwhile, mostly open source. I am a co-owner of a company that sells both: SaaS licenses and self-hosting support. After 15+ years in this business I know the trade-offs of both. But if you ask me what I would use for anything business-critical there is only one answer: self-hosted.

When to Choose SaaS in Digital Signage

There are some honest reasons for a SaaS solution.

• you’re running a handful of screens for a short-term campaign
• you just need something up and running without the hassle
• you have no sensitive data and no long-term commitment
• you want to test or evaluate a business idea
• you do not want for no reasons to babysit a server

Identity management is a good example. Using an existing SSO with a provider like Azure AD or Okta means your team can log in from day one without setting up anything.

Some vendors offer hybrid models. Same software, but deployable on-premise. This can be worth considering if you want control without building software on your own. It also gives you some security if the vendor disappears or changes their strategy.

SaaS is just someone else’s computer. Nothing wrong with that. But almost nobody selling it will tell you what that actually means long-term.

When It Starts to Become Uncomfortable

The SSO example works fine until your provider changes pricing, gets acquired, or your account gets suspended and your whole team is locked out. The subscription that felt cheap in year one looks different in year three. Your screens depend on decisions made in a boardroom you have no access to. Your data sits on infrastructure you don’t control, often in jurisdictions with different privacy laws than yours. And when the provider has an outage, you just wait.

As an EU-based business, you’re responsible for your users’ data regardless of where it’s stored. GDPR compliance is your obligation, not your SaaS vendor’s.

This isn’t a digital signage problem. This is a common SaaS and cloud problem. Broadcom acquired VMware and overnight turned reasonable licensing into something many companies simply couldn’t afford anymore. Microsoft can lock your entire organization out of Teams and Exchange the moment a license lapses, payment fails, or a political decision is made.

Twitter killed its API and took dozens of businesses that had built on it down with it. Google has a graveyard of discontinued products that people and companies relied on. Heroku killed its free tier and thousands of small projects went dark the same day. Someone always makes the decision for you.

In digital signage the consequences are just more visible. Your screens go black or show some errors. There is no workaround and no fallback. You wait for someone else to fix their problem.

Self-Hosting

Self-hosting means you run the software on your own infrastructure. Your server means your data. No one can double your prices, kill the product, or lock you out. You decide when to update, whether to update, and what to do with your own data.

Open Source for More Transparency

Open source makes things much more interesting. The code is auditable, you’re not trusting a black box. If I stop developing GarlicSignage tomorrow because I choose to continue my life as a Buddhist monk, the code stays. Someone can fork it and maintain it. The community doesn’t depend on my business decisions. That’s worth more than any SLA.

For regulated environments like healthcare or the public sector, auditability is often a hard requirement. You can’t put a closed-source black box on a hospital network and call it compliant.

Remember the hybrid model? At least it lets you run the same software on your own infrastructure. But without open code you’re still trusting a black box. You cannot audit or fork it. And there is no exit if the vendor changes the terms.

Self-hosting Can be Tricky.

I’m not religious about it. Someone has to set it up, keep it running, and fix things when something breaks. Most vendors sell you the advantages of their prefered model and stay quiet about the rest.

For a small business with no technical background, “just install a server” is not really helpful. In 2026 that usually means “just run docker compose”, which lowers the bar significantly, but you still need someone who knows what they’re doing. That might be an employee, a freelancer, or a managed hosting provider. Either way, it’s a dependency of a different kind, one you at least have some control over.

You have to care about updates. That’s a feature if you value stability and hate surprises. If you ignore them for too long you will run into problem sooner or later. Self-hosting needs continuously care. Backups and Monitoring are also your responsibility. If your server goes down at 3am, no one is going to fix it unless you or someone you hired does.

And then there’s the setup itself. Depending on what you’re running, it can be straightforward or a rabbit hole. You are lucky if you find a well-documented open source project with an active community. But it can be also poorly maintained one with no docs.

Some people romanticize full control without thinking about what it actually costs them in time and attention. That’s how they end up abandoning it halfway and blaming the software.

Let’s Make it Simple

SaaS costs less when you start and more if you want to leave. Self-hosting costs more time and attention, but pays off the longer you run it.

I offer both models because both have their place. But if you ask me what I’d run my own critical infrastructure on: My first choice is mostly open source and self-hosted. Not because I’m ideological about it, but because I’ve watched too many businesses trapped in a dependency trap and struggle with the decision others made for them.

Google discontinued AngularJS in favor of a TypeScript-based Angular. A friend of mine had to hire people and spend serious money migrating his codebase. His business, his decision. I could not understand that and would never rewrite my codebase with a product from a company that already burned me once.

My skepticism about depending on other people’s software decisions goes back further than that. In 1992, I was an enthusiastic OS/2 user. Maybe even a little bit too much. Fact: OS/2 was technically superior to Windows and had great object-oriented GUI concept. But it was backed by IBM and then slowly strangled. Partly because it wasn’t profitable enough, partly because Microsoft made sure it wouldn’t be. Everyone who had built software or a business model on it had to start over. That stuck with me. It might be worth thinking about what’s going to stick with you.