Software as a Service

In the past, you bought programs in a box with a CD and installed them on your own computer. With Software as a Service, the program instead runs on powerful computers belonging to the provider, which are permanently connected to the internet. You simply access it in the browser and work there. Payment is not made once, but regularly – like a magazine subscription. The English name literally means “software as a service”: you’re not buying a product, you’re buying the usage. Well-known examples are Google Docs, Microsoft 365, Spotify, and Netflix.

Why investors watch subscription revenue

SaaS is one of the most common terms in business news about technology companies. The reason is the business model. A company that sells software once has to find new buyers every month. A company with subscriptions, on the other hand, knows fairly precisely how much money will come in next month. This predictability is considered very valuable on the stock market.

That’s why certain metrics show up with such companies. The most important is Annual Recurring Revenue: the sum of all ongoing subscriptions, calculated on an annual basis. A second one is the churn rate, often simply called “churn.” It indicates what share of customers cancel per year. If it’s low, the company grows almost automatically.

For customers, the model has advantages and disadvantages. Advantage: no high upfront costs, no installation, always the current version. Disadvantage: you pay continuously and are dependent on the provider. If they raise the price or discontinue the service, you have little room to maneuver. Experts call this dependency the lock-in effect.

One program, many tenants

Technically, behind every SaaS offering is a data center: a hall full of servers, i.e. computers that answer requests around the clock. Your browser sends a request there, the server computes, the result comes back. Your laptop does hardly any work in the process. That’s why Google Docs also works on an old, slow device.

A principle called multi-tenancy is decisive for the economics. It means: all customers use the same single version of the program, only their data is strictly separated from each other. You can think of it like an apartment building. There’s one building, one heating system, one caretaker – but every apartment has its own lock. That is exactly where the cost advantage comes from: an update only needs to be installed once and immediately applies to millions of users.

SaaS is often equated with “cloud,” but that’s imprecise. Cloud is the umbrella term for rented computing power. SaaS is the tier at which you rent a finished program. If you only rent virtual computers without programs on them, that’s called Infrastructure as a Service. SaaS is thus the most customer-friendly, most fully prepared variant.

From Spotify to the AI assistant on subscription

In everyday life, almost everyone uses SaaS daily without knowing the term. Writing a report in Google Docs, storing photos in iCloud, streaming music: all software that runs elsewhere. School platforms and video conferencing services like Zoom also belong to this category.

In companies, the model is even more widespread. Customer management, accounting, human resources – for every department there are providers who sell their software as a subscription. Salesforce is regarded as the pioneer of this industry and has sold no installation files but only access since the late 1990s.

Currently, SaaS is meeting artificial intelligence. Services like ChatGPT, image generators, or coding assistants are sold exactly according to this pattern: monthly fee, use in the browser. For providers, this is trickier than with classic software. Every response from an AI model costs computing power and thus real money. With a word processor, a heavy user hardly causes extra costs; with an AI chat, they do. That’s why pricing models with usage limits instead of a true flat rate are increasingly seen there.