Agencies love time & materials, clients love fixed prices — and both sides have their reasons. The real question is not “which model”, it is “who carries the risk in it”.
When a fixed price makes sense
When the scope is definable: an MVP, a website, an integration with a known API. The fairness condition is a paid discovery phase — without it, the vendor simply prices the risk into a 30 % buffer and you pay it anyway, just invisibly. A fixed price with defined KPIs and a delay penalty is the strongest client protection on the market.
When time & materials is fair
For long-term product development where priorities shift monthly. Fairness conditions: weekly demos, a transparent hourly statement, and the right to end with one month's notice. If a vendor on T&M refuses measurable weekly outputs, you are not paying for development — you are paying for a promise.
Change requests: where a fair model turns into a trap
With a fixed price, the most money is lost not in the price but in change requests. If discovery is shallow, half the real work “was not in scope” and gets billed hourly — suddenly the fixed price is not fixed. So ask up front for the change-request rate and, above all, a proper specification out of discovery. The more detailed the scope at the start, the fewer surprises on the invoice.
The hybrid model that protects both sides
The healthiest combination separates the known from the unknown. The phase you can describe (an MVP, an integration) is fixed; ongoing development, where priorities shift, runs on T&M with a monthly cap and the right to stop. The client never pays blind and the vendor never works for free.
Our model: fixed discovery, fixed MVP, then T&M with a monthly cap. Risk sits with whoever can manage it — scope with us, priorities with you.