Agency, freelancer or platform
All three build the same app and sell different things. Choosing on price is the fastest way to choose wrong, because what differs between them is not only the cost.
Agency
You are buying process and accountability. There is somebody to hold responsible, a contract, a team that still exists if one person leaves.
It wins when: the project is large, has a contractual deadline, integrates with an existing system, or when the company needs a supplier with a legal entity and an SLA.
It loses when: the app is small. There is a floor of hours that does not shrink (discovery, design, testing, delivery), which is why an agency quote for a menu app looks disproportionate. It is not overpriced; it is at its own minimum.
Freelancer
You are buying one person's hours. It scales down better than an agency, and it is where a small app genuinely costs less when somebody has to build it for you.
It wins when: the scope is clear, you know what you want, and having someone dedicated outweighs the risk.
It loses when: the person disappears. The real risk is not quality, it is continuity, and it drops sharply with one simple precaution: the code and the store accounts in your name from day one.
Platform
You are buying access to a tool and building it yourself. The price is fixed, monthly, known before you start, and does not grow with the scope. At Fabapp: Free to build and publish on the web, paid plans from R$ 99 a month.
It wins when: you want to start today, change your mind next week without renegotiating, or build several apps for the same price.
It loses when: nobody has time to build. The platform reduces the cost of the work; it does not remove the work.
The four questions that separate proposals
- who keeps the code?
- whose name are the Google and Apple accounts in?
- what does the second year cost?
- what happens if you disappear?
The fourth is the most awkward and the most revealing. The first three have objective answers, and a proposal that dodges any of them has already answered.
The combination that is usually the best deal
Build on the platform, publish, find out what people actually use. If a part shows up that the tool cannot handle, export the code and hire someone for that part only.
You pay for hours only where hours are needed, and you arrive at the conversation with a working app, which is the best possible position from which to hire. Without export that combination does not exist, which is why question 1 is first on the list.
The summary
- an agency sells process, a freelancer sells hours, a platform sells a tool;
- a small app is where the agency saves you least, because of the floor of hours;
- the freelancer risk is continuity, and it drops with code and accounts in your name;
- ask about the second year, not the first;
- building on a platform and hiring only for the hard part is possible only with export.