A Paid Discovery Phase: What You Should Receive Before Committing to a Software Build
November 11, 2026 · 7 min read · Etzal Group Team
A company asks three vendors to quote for "a field service management system". The quotations come back at 400 million, 950 million and 2.1 billion rupiah. None of the vendors is dishonest. Each imagined a different system, because the request described a title and not a product. The company picks the middle one, and four months in, both sides discover what they had each assumed. A discovery phase is how a business finds out what it is actually buying before it signs for the build.
The problem discovery solves
At the start of a project, nobody knows enough.
- The client knows the business and cannot yet say precisely what the software must do.
- The vendor knows software and does not yet understand the business.
- Any price given at this point is a guess, padded for uncertainty or dangerously optimistic.
Discovery spends a little time and money to remove most of that uncertainty. The build that follows can then be priced, planned and judged against something concrete.
What you should receive
A proper discovery produces documents you can hold. Seven are typical.
1. A problem statement and goals
One or two pages. The situation today, what it costs, and what success looks like in measurable terms. "Reduce the time from job completion to invoice from five days to one."
2. Process maps
How the work flows today and how it should flow with the system. Who does what, in what order, with what information. This is often the first time anyone has drawn it, and the drawing alone reveals problems that need no software.
3. Requirements, prioritised
A list of what the system must do, written as user needs, each marked:
- Must have for the first release.
- Should have soon after.
- Could have later.
- Will not have, stated explicitly.
Plus the non-functional requirements: number of users, data volume, response time, availability, security, devices.
4. A clickable prototype
Screens for the main journeys that users can click through. No real data behind them, and that is enough: people react to what they can see. A prototype shown to real users in week two prevents months of rework.
5. An architecture outline
In plain terms: the main components, how they connect, what existing systems are integrated and how, where data lives, the technology proposed and why. Enough for a second opinion.
6. A delivery plan
Phases and releases, what each contains, dependencies on the client, and the main risks with how each will be handled.
7. A firm quotation for the build
A price for the first release, tied to the requirements and plan, with a breakdown, assumptions and exclusions. Also an estimate of running costs.
You should own all seven. The contract for discovery should say that the deliverables belong to you and can be given to any other vendor.
What a typical discovery looks like
For a mid-sized business system, two to four weeks.
Week 1: understand.
- Kickoff with the sponsor: goals, constraints, budget range, decision makers.
- Interviews with people in each role, including those who do the daily work.
- Observation: watch the work being done, look at the spreadsheets and forms in use.
- Review of existing systems and data.
Week 2: shape.
- Process maps, current and future.
- First requirements list.
- Sketches of key screens, reviewed with users.
Week 3: test.
- Clickable prototype of main journeys.
- Sessions with real users: can they complete the task? What is missing?
- Technical investigation of the riskiest points: an integration, a data migration, a performance question.
- Architecture outline.
Week 4: decide.
- Prioritisation workshop: what is in the first release.
- Plan and quotation.
- Presentation to the sponsor, and a decision.
Smaller projects compress this into one or two weeks. Larger ones extend it.
Who must be involved
Discovery fails when it is done to a client instead of with them.
- The sponsor, who owns the budget and the outcome. At the start and the end, and available for decisions between.
- People who do the work. Not only their managers. The clerk who keys the orders knows where the process really breaks.
- Someone from finance and from IT, where data and systems are touched.
- One product owner on the client side, who can answer questions daily and resolve disagreements between departments.
Budget the client's time, not only the vendor's. A discovery where users could spare only an hour produces a document about the managers' assumptions.
Signs of a real discovery
- They ask more than they tell.
- They talk to users, not only to the sponsor.
- They watch the work.
- They tell you things you did not know about your own process.
- They challenge the brief. "You asked for a mobile app. From what we saw, the real problem is the approval step."
- They say what should not be built.
- They investigate the risky parts instead of assuming.
- The prototype is tested with users.
- The quotation has clear exclusions and assumptions.
Signs of a sales exercise
- A generic document with your logo on it.
- Requirements that restate your original brief.
- No conversations with actual users.
- No prototype.
- A quotation that arrives suspiciously quickly and matches your stated budget exactly.
- Deliverables you may not take elsewhere.
What can come out of it
Discovery ends with a decision, and four outcomes are all legitimate.
- Build as proposed. The usual result.
- Build something smaller. The first release is cut to what matters most.
- Do not build: buy. A ready-made product turns out to cover the need.
- Do not build yet. The process itself must be fixed first, or the data is not ready.
A discovery that concludes "do not build this" has saved its fee many times over. Be suspicious of a process that could never reach that conclusion.
How to pay for it
- Fixed price for a defined set of deliverables and a fixed period.
- Separate contract from the build. You are free to stop after discovery.
- Ownership of deliverables with the client.
- Optionally, a credit: some vendors deduct part of the discovery fee from the build if you proceed. Pleasant, and it should not influence the decision.
As a rough guide, discovery commonly costs in the region of five to ten percent of the build that follows. The percentage is higher for small projects and for problems that are poorly understood.
When you can skip or shorten it
- The system is small and simple.
- You already have good requirements, process maps and a prototype.
- You are extending a system the vendor already knows.
- You are working in short paid iterations with frequent review, where each iteration is a small discovery in itself.
Even then, a few days to confirm understanding before a fixed price is given is worth having.
After discovery
- Use the documents. The requirements become the scope in the build contract. The prototype becomes the reference for acceptance.
- Expect change anyway. Discovery reduces surprises. It does not end them. Keep a change process.
- Keep the users involved. The people who tested the prototype should see the working software every few weeks.
Common mistakes
- Asking for a fixed quotation with no discovery.
- Treating discovery as a formality before a decision already made.
- Sponsor present, users absent.
- No prototype.
- No investigation of technical risks.
- A wish list with no priorities.
- Deliverables that stay with the vendor.
- Refusing to accept "do not build" as an answer.
We start most projects with a fixed-price discovery, and clients keep everything it produces whether or not they continue with us. See how we scope work in Enterprise Software, or contact us to discuss what a discovery for your project would cover.
Summary
Before committing to a build, pay for two to four weeks of discovery and expect seven things: a problem statement, process maps, prioritised requirements, a tested prototype, an architecture outline, a plan and a firm quotation. Involve the sponsor, a product owner and the people who do the work. Own the deliverables, keep the contract separate from the build, and accept that the right conclusion may be to build less, buy instead or wait.
Frequently asked questions
What is a discovery phase in a software project?
It is a short, paid piece of work before the build, in which the vendor and client study the problem together and produce the documents needed to build with confidence: requirements, process maps, a prototype, an architecture outline, a plan and a firm quotation. It replaces guessing with evidence.
Why pay for discovery when some vendors offer it free?
Free discovery is paid for by the vendor's sales budget, which limits its depth and ties it to winning the build. Paid discovery gives you the vendor's proper attention and deliverables you own outright, which you can use to obtain quotations from anyone. It also tests the working relationship at low cost.
How long does discovery take and what does it cost?
For a mid-sized business system, two to four weeks is typical. As a rough guide, discovery often costs in the range of five to ten percent of the eventual build, though it varies with complexity and how much is already documented. Measured against the cost of building the wrong thing, it is small.
Related service
Free consultation
Not sure where to start? We will map the right system for your business, free.
