Self-Hosting vs. SaaS: The False Choice

The Same Conversation, Every Time

It starts with a tool evaluation. The team needs a database, a queue, an auth service, an analytics pipeline. Someone suggests a SaaS option — fast to set up, someone else’s problem to run. Someone else points out the cost at scale, the data residency issue, the fact that if the vendor raises prices or gets acquired, you are stuck.

So the conversation pivots to self-hosting. Run it yourself. Full control. Then the operational questions arrive: who maintains it? Who applies security patches? Who is on call when it falls over at 2am? Who manages the upgrades that break backwards compatibility?

The meeting ends without a decision, or with a decision that will be revisited in twelve months when the pain from whichever choice was made becomes undeniable.

This conversation happens in almost every engineering organisation. It never fully resolves because both options are genuinely bad in different ways.

What SaaS Actually Costs

The pitch for SaaS is compelling: someone else handles the infrastructure, the patching, the scaling, the monitoring. You pay a subscription and get a working service. For early-stage teams or non-core tooling, this is often the right call.

But the cost structure of SaaS is not just the invoice.

Control. Your data lives in someone else’s infrastructure, under their security model, in jurisdictions they choose. For many organisations — regulated industries, European companies under GDPR, public sector, defence — this is not a trade-off they can make regardless of convenience.

Price at scale. SaaS pricing is designed to be cheap when you are small and expensive when you are not. The introductory tier that made the tool easy to adopt becomes the pricing lever once switching costs are high enough. Many organisations have found that a SaaS line item that seemed trivial at adoption has become one of their largest infrastructure costs two years later.

Vendor dependency. The more deeply a SaaS tool integrates into your workflows and data model, the more expensive it becomes to leave. This is by design. Integrations, proprietary APIs, exported data formats that are not quite standard — these are switching costs that compound over time.

Availability on someone else’s terms. Outages, deprecations, feature removals, sunset notices — these are events in someone else’s roadmap that land in your incident log. You have no control over when they happen or how much warning you get.

What Self-Hosting Actually Costs

The pitch for self-hosting is equally compelling: you own the infrastructure, the data stays where you put it, and you are never at the mercy of a vendor’s pricing decisions.

The costs here are less visible but just as real.

Operational burden. Running software in production is a job. Patching, upgrading, monitoring, backup, recovery, capacity planning — someone has to do all of this for every piece of infrastructure you self-host. For teams without a dedicated platform function, this work competes directly with product development.

Security surface. Self-hosted software is your responsibility to keep secure. A misconfigured database, an unpatched vulnerability, a container running with too many privileges — the SaaS vendor would have caught and fixed these silently. Self-hosted, they are your exposure until someone on your team notices.

Expertise tax. Operating Postgres well is not the same as using Postgres. Running Kafka at scale is not the same as publishing to a Kafka topic. Self-hosting specialist infrastructure requires specialist knowledge your team may not have and may not want to develop.

Hidden total cost. When organisations do the maths on self-hosting — including engineering time, on-call burden, security reviews, and the opportunity cost of not building product — the cost advantage over SaaS often narrows or disappears entirely.

The Real Question

The SaaS vs. self-hosting debate is usually framed as a convenience-versus-control trade-off. But that framing assumes the choice is binary, and it does not have to be.

The actual question is: who controls the operational layer, and at what cost?

SaaS answers this by putting the vendor in full control. Self-hosting answers it by putting your team in full control at full cost. Neither is satisfying because neither is actually what most organisations want.

What they want is to own their data and their infrastructure decisions — without having to build and operate everything from scratch themselves.

The Gap in the Tooling

This gap exists because running software in production has historically required either specialist knowledge or complete delegation to a vendor. There is no middle ground — no way to deploy a CRM, an email platform, or a document collaboration suite with the same ease as a SaaS sign-up, but on your own infrastructure, under your own control.

That is what makes the SaaS trade-off feel necessary even when organisations know it is the wrong long-term choice.

The emerging answer is not better documentation for self-hosted software, and not cheaper SaaS pricing. It is a different model altogether: a marketplace of applications you deploy into your own environment, rather than rent from someone else’s.

Think of the tools your organisation pays SaaS subscriptions for today — CRM, email, project management, document collaboration, customer support, HR software. Each of these exists as open-source or self-hostable software. The reason most organisations use the SaaS version is not that it is fundamentally better. It is that deploying and operating it themselves has historically cost more in time and expertise than the subscription price.

A marketplace changes that equation. Pick an application, deploy it into your own environment in minutes, and get something that is already secured, already configured, and maintained through the platform going forward. You get the convenience of SaaS with the control and economics of running it yourself — not as a compromise, but because the operational expertise is built into the deployment rather than left for your team to figure out.

The choice between SaaS and self-hosting is not a law of nature. It is a gap in the tooling, and gaps get filled.