There are two ways to hire custom software development: hourly, where you pay for time worked, or fixed price, where you pay a set amount for a defined scope. For a small or medium business that needs a concrete problem solved, fixed price is almost always the better deal, because the risk of the project dragging on sits with whoever builds it, not with whoever pays for it. Hourly makes sense when nobody knows yet what needs to be built.
The rest of this article explains why, when it doesn't apply, and what to check in a proposal before signing it.
What's the difference between fixed price and hourly?
The real difference isn't how the number is calculated. It's who pays if the project takes longer than planned.
| Fixed price | Hourly | |
|---|---|---|
| What you pay for | A result: the working system | Working time |
| Final cost | Known before starting | Known when it's over |
| If it runs late | The provider absorbs it | The client pays for it |
| Scope changes | Quoted separately, as a new phase | Added to the bill with no friction |
| Requires | A scope defined in writing | Trust and constant follow-up |
Why does hourly pricing usually end up costing more than expected?
Because the incentive points the wrong way. Someone billing by the hour gains nothing from finishing early: every extra week is revenue. It doesn't take bad faith for that to matter. It's enough that nobody has a strong reason to cut a meeting short, simplify a feature, or say "we don't need this."
The second problem is that an hourly rate doesn't tell you what the project will cost. A cheap hour from someone slow costs more than an expensive hour from someone fast, and from the outside you can't tell which one you're hiring. The initial budget "in estimated hours" commits no one: it's an estimate, and software estimates fall short far more often than they overshoot.
What's the risk with fixed price?
A poorly defined scope. If the proposal says "customer management system" and nothing else, each side will understand something different, and the argument shows up when the project is already well underway. A fixed price on a vague scope is worse than hourly, because it gives a false sense of certainty.
The other risk is the provider who quotes low to win the project and then makes up the difference by billing as "extras" things that were obviously part of the job. You can spot it before signing: if the proposal doesn't list what's included and what isn't, that room for interpretation exists and someone will use it.
When does hourly make sense?
- When you don't know yet what to build. Exploration, prototypes, researching whether something is feasible. You can't fix a price on something nobody can describe.
- Maintenance and small changes to an existing system. One-off tweaks that don't justify writing a proposal every time.
- When you're adding a person to your own technical team. There you're hiring capacity, not a result, and you direct the work.
Outside those cases, if the problem can be described, it can be quoted at a fixed price.
What should a fixed-price proposal say?
Before signing, the proposal should answer these six questions in writing:
- What gets built, feature by feature, not in general headings.
- What doesn't get built. The exclusions list matters as much as the inclusions.
- In how long, with partial deliveries and not a single date at the end.
- How it's paid and against which deliveries.
- What happens with changes: how something that comes up halfway gets quoted.
- What costs continue afterward: hosting, maintenance, licenses, and who owns the code.
If any of the six is missing, ask. The answer says a lot about what working with that provider will be like.
How do we work at RUBRA lab?
Fixed price. Before writing a line of code we define what gets built, in how long, and at what price, and that number doesn't change halfway through. Projects start at USD 1,000 and most are delivered in two to six weeks, in phases: there's something working to see and correct from the start, not a single delivery at the end.
Whatever comes up outside the scope isn't argued over as a rushed extra: it's quoted as a new phase, with its own fixed price. After launch, hosting runs about USD 25 per month, support is optional starting at USD 75 per month, and the code is 100% the client's, with no licenses and no dependency on a platform of ours.