A fixed price is the most requested and least understood way to buy software. Clients ask for it because they want a number they can plan around. What they often miss is that the number is fixed against a defined scope, not against the vaguer thing they actually have in their head. The claim worth defending here is simple: a fixed price prices the effort, not the discovery — and if discovery has not happened yet, the price is a guess wearing a suit.
What a fixed price actually prices
Fixed pricing works when scope is knowable in advance: a defined set of screens, a documented integration, a migration with a fixed number of records. The vendor absorbs the risk of how long the work takes, which is a genuine and valuable transfer of risk. What it cannot absorb is the risk of what the work is changing, because that was never in the contract.
The four things that reliably break a fixed price
In practice, fixed-price projects overrun for the same four reasons almost every time: requirements discovered mid-build that were assumed rather than stated; third-party systems that behave differently from their documentation; data that is messier than the sample provided; and stakeholders who were not in the room when scope was agreed. None of these are the developer being slow. They are scope arriving late.
A worked example
Consider a quote of ₹8,00,000 for a portal estimated at 40 developer-days at ₹20,000/day. If discovery reveals the client's legacy export is CSV with inconsistent date formats across three source systems, cleaning and reconciling that data might add 8 days. On time-and-materials that is a ₹1,60,000 line item the client sees and approves. On a fixed price it is either an awkward change request or margin the vendor quietly eats — and eaten margin is where quality goes to die.
When fixed price beats time-and-materials
Fixed price is the right instrument when the work is well-bounded and the client values a guaranteed ceiling more than flexibility: a marketing site, a defined API integration, a phase two with a phase one already built. Time-and-materials fits genuine discovery, where the honest answer is that neither side yet knows the full shape of the work. The mistake is forcing an exploratory project into a fixed number to feel safe, then discovering the safety was imaginary.
What to ask any vendor before signing
Ask three questions. What exactly is in scope, written as a list you could tick off? What happens when something outside that list appears — the change process, in writing? And how much discovery has actually been done to produce this number? A vendor who has run a short paid discovery before quoting is pricing reality. A vendor who quoted from a one-page brief is pricing hope.
We scope fixed-price work off a short discovery phase precisely so the ceiling means something — you can see how we approach that on our product development service.