Paying before the building starts feels backwards to most owners, so discovery gets skipped, and that skip is the most expensive line item in software. Here is what discovery is, what it costs, and why it is the highest-leverage money in the whole project.
What discovery actually is
A one to two week paid engagement that answers, in writing: what are we building, for whom, on what technology, in what order, for how much, and by when. It ends with artifacts, not meetings:
- A scope document listing every feature and, critically, what is excluded
- A sitemap or screen list (and rough screens if design follows)
- Technical decisions made deliberately rather than by habit
- A timeline with milestones
- A fixed price for the build, based on the plan instead of a hunch
Why it costs what it costs
Discovery is real specialist work: interrogating requirements, mapping edge cases, checking integrations, and designing the plan. A developer doing discovery properly is spending focused senior hours, and the artifact they produce is the thing every later hour depends on.
Typical pricing: $500 to $1,500 for a small business site, $1,500 to $3,000 for an app or web app with accounts and payments. Many developers, me included, credit it against the build if you proceed. That structure tells you something: developers confident in their process can afford to make discovery free-on-continuation, because planned projects go well.
The math that makes it obvious
Say discovery costs $1,000 and prevents one mid-build misunderstanding: the "oh, I assumed checkout was included" kind. That misunderstanding, discovered in week 6, costs a change request ($500 to $2,000), a timeline slip (your launch moves), and trust (both sides now document everything defensively, which slows the rest).
Discovery does more than prevent one misunderstanding; it removes the whole class. Quotes written after discovery are tighter, builds go faster because decisions are pre-made, and the end-of-project negotiation, the ugliest part of freelancing, mostly cannot happen because the finish line is written down.
When you can honestly skip it
- You can point at an existing thing and say "that, for my business" (a landing page, a clone with known scope)
- The job is under a couple thousand dollars
- You are using an established platform with standard patterns and no integrations
Skipping discovery on a $1,800 brochure site is sensible. Skipping it on a $15,000 booking platform is how $15,000 becomes $22,000 with a broken launch.
What a good discovery phase looks like week by week
Week 1: interviews with you (and your staff who will use the thing), review of existing systems, competitor scanning, technical spikes on anything risky like an odd integration.
Week 2: the writing: scope, exclusions, screens, plan, fixed quote, and a walkthrough call where you challenge everything. You approve or push back; the document updates; then both sides sign.
You should end this phase able to hand the document to any competent developer and get consistent build quotes. That portability is also your proof the discovery was honest and not a setup for a lock-in.
The question to ask any developer who skips discovery
"Then how do you know what to build?" If the answer is "we figure it out as we go," you are funding the discovery anyway, one confusing invoice at a time.
If you have a project in mind and are not sure whether it needs discovery or can go straight to a build, describe it to me in a few sentences. I will tell you honestly which one it is, and what discovery would cost for it if it does.