Software house, software factory, agency or freelancer: which solves your problem

When you decide to build something digital, you discover that there are five types of suppliers offering you the same thing under different names. Software house, software factory, digital agency, consultancy, freelancer. Each swears they do what you need. And, in part, everyone does. The problem is that they do it in such different ways that choosing the wrong model is expensive even when the supplier is good.

The confusion is intentional in some cases and innocent in others. I will separate the models by what matters to you: how each one thinks, what each one is good at and what type of problem each one solves without letting you down.

Software house: the construction partner

The software house positions itself as someone who builds custom software and participates in decisions. She doesn't just want to receive the specification and execute it. You want to understand the problem, give your opinion on the path and be responsible for the result over time.

It's the right model when what you need is specific, will evolve and has software at the center. A new digital product, an internal system that supports the operation, a complex integration between things that don't talk to each other. Cases in which asking for the right thing is difficult and having someone thinking together is worth more than the price per hour.

The cost of this model is what he discusses. A good software house will question your request, and this conflicts with those who just want it to be executed. But it is precisely friction that prevents building the wrong thing. If you want to better understand the concept, I wrote an article just about what a software house is.

Software factory: running at scale

The software factory is optimized to execute specifications with predictability and volume. You arrive with the defined scope, it transforms it into code efficiently. The name "factory" is honest: the strength is the production line, not the questioning.

This is great when the problem is already well understood and the specification is solid. Companies that already have an internal product team, that know exactly what they want and need the ability to execute, do very well with factories. You bring the brain, she brings the arms.

The risk appears when you don't know what you want and hire a factory expecting guidance. She will execute what you asked, faithfully, including the mistakes. And it will charge you again for a redo when you realize you ordered it wrong. Brainless factory on the outside becomes a machine for producing the wrong thing efficiently.

Digital agency: the showcase

The digital agency is born from marketing and communication. DNA is the appearance, the brand, the experience of those who visit. Many agencies do development, and some do it well, but the center of gravity is visuals and campaigns, not engineering.

For an institutional website, a landing page, a well-designed digital presence, the agency is usually the right choice and cheaper than a software house. This is her land.

The problem is when the project has software complexity underneath (business rules, integrations, robust logged area, scale) and the agency accepts the work anyway. The result tends to be beautiful on the outside and fragile on the inside. When the system grows, there is a lack of engineering, and you discover it late. Evaluate the agency by what it really masters, not by the prettiest portfolio.

Consulting, freelancing and SaaS ready

Consulting sells decisions, not execution. You hire her when the problem is "what should I do" before "who will build it." Technology strategy, architecture, choice of path. Good consultancy is worth the price when the decision is expensive and you don't have anyone to make it internally. But consultancy that doesn't execute leaves you with a beautiful report and the problem of finding someone to build it.

Freelancing is the cheapest and riskiest option. For something small, punctual and well-defined, a good freelancer can solve it quickly and at a low cost. For something that needs to last, the risk is structural: a person just gets sick, disappears, changes priority, and your project stops. It's not a question of individual competence, it's a question of there being no structure behind it.

Ready-made SaaS is what a lot of people forget to consider. Before building anything, ask if there isn't already a tool that solves the problem. If it exists and serves the purpose, building to measure means spending more to have less and still inherit the maintenance. Serious software house will tell you this, even losing the sale.

How to choose between them

The question that organizes everything is: is your problem well understood or does it still need to be discovered? If it is clear and stable, a factory or freelancer will do it cheaper. If it's still nebulous and the software is central, you need someone who thinks together, and that's where the software house comes in.

The second question is about durability. Does what you are going to build need to live for years and evolve, or is it a one-off? The longer the software needs to last, the more you need structure behind it, and the less a freelancer or agency without engineering will serve you.

The third question is about who will think about the project. Do you have someone in-house who defines what to build and covers quality, or do you need the supplier to bring that brain? If you have your own management, an efficient factory is great and more economical. If you need someone thinking together, paying for a cheap executor is expensive, because no one will stop you from building the wrong thing.