Fixed price works when the outcome is one workflow you can demo and accept. Time and materials works when discovery is still open and every week changes the backlog. Mixing the two without saying so is how founders get surprise invoices.

A clean pattern: pay for a short discovery that produces scope, risks, and acceptance criteria. Then lock a fixed-price build for that slice only. If the product starts needing continuous iteration after launch, switch the engagement model deliberately instead of pretending the original SOW still fits.

Ascendra Ventures prices most early builds as one atomic technology problem with a fixed outcome, because founders usually need a shippable chapter more than an open-ended team. Worth using that split when you evaluate the models in this post.