How Much Does a Discovery Phase Cost? What You Get and Why It Pays Off

A discovery phase typically costs €1,500-5,000 with an experienced European freelance developer, which buys 15-50 hours of focused work, and up to around €10,000 for a large platform with many integrations. A useful rule of thumb is that discovery phase cost should be about 5-10% of the expected build budget, and it pays for itself the first time it catches an expensive misunderstanding before any code is written.

I sell fixed-price discovery phases myself, so weigh my view accordingly. That's also why there's a section below on when you should skip discovery altogether.

The short answer: what discovery costs by project size

The size and fuzziness of the final build decide how much discovery you need. These are my rough estimates for an experienced freelance developer, calculated at €100 an hour excluding VAT and rounded.

I use €100 an hour because it's a round number close to the European freelance market: the freelancermap 2026 market study, based on more than 5,000 IT freelancers in Germany, Austria and Switzerland, puts the average at €103 an hour and development roles at €91. Senior freelancers in Denmark sit higher: LønRadar's 2026 guide puts senior IT consultants at 800-1,200 DKK an hour, roughly €105-160. Treat €100 as an input you can swap for a real quote, not as my rate.

Agencies usually land above these figures. The hourly rate is often higher, but the bigger factor is headcount: when a project manager, a designer and a developer all join each workshop, every hour is billed for three people.

Discovery is one line in a much bigger budget. For what the build itself tends to cost, see my guide to software development costs.

What you should walk away with

A discovery phase is only worth the money if you end up with something concrete. Before you sign, check that the proposal promises these seven things:

A slide deck full of personas and sticky-note photos is not the same thing. The test is simple: could a developer who wasn't in the room pick up the material and start building, or price the work, without guessing?

The output overlaps with a requirements specification, with one difference that matters: it's written together with the person who has to estimate it. If you'd like to prepare some of it yourself, my requirements specification template is a good starting point.

How a discovery phase runs

With a freelancer, discovery typically takes one to three weeks of calendar time, and your availability for meetings and feedback usually sets the pace. Workshops work fine over video, and a developer in Central European Time overlaps almost completely with UK and EU office hours.

1. Kickoff. You walk through the idea, the goals, the users and the systems you use today, including a rough budget and timeline.
2. Research. Existing tools, spreadsheets, data and integration docs get reviewed so the plan rests on reality.
3. Wireframes and priorities. The key screens are sketched and the feature list is sorted into now and later.
4. Technical check. The data model, architecture and integrations are settled, and a missing API (the connection point other software talks to) surfaces now rather than in month three.
5. Estimate and review. Hours are split by phase, risks are written down, and you decide whether to build, cut back, wait or stop.

I run a discovery phase myself before larger builds. It has a fixed price you know before work starts, and you work directly with me, the person who will write the code. So the estimate comes from the same person who later has to deliver on it.

Why discovery usually pays for itself

An estimate based on a loose idea is an educated guess. Barry Boehm described back in 1981 how uncertainty in software estimates is highest at the start and shrinks as decisions get made. Steve McConnell later popularized it as the cone of uncertainty, and in the usual presentation of the model, an estimate made at the concept stage can be off by a factor of four in either direction. Discovery is where you pay a small amount to narrow that range before the big money is spent.

A worked example

The numbers below are made up, but the pattern is common. Picture a B2B SaaS product estimated at 400 hours, or €40,000 at €100 an hour. Discovery takes 30 hours and costs €3,000, which is 7.5% of the budget.

Your project won't hit all three. But each one costs about as much as the discovery itself or more, so avoiding a single one roughly covers the fee.

The benefits that don't show up in the sum

• A tighter quote. With a clear scope, a developer doesn't need a large buffer for the unknown, so a fixed price for the build becomes realistic.
• Quotes you can actually compare. Send the same material to two or three developers and you're comparing prices for the same product. Here's how to compare software development quotes side by side.
• A trial run of the relationship. After a couple of weeks you know how the developer communicates, before you commit tens of thousands of euros.
• A cheap way out. As the GOV.UK Service Manual puts it, "It's not a failure to stop at the end of the discovery phase" when the research points that way. Stopping after €3,000 beats stopping after €30,000.

What makes discovery cost more or less

The hourly rate is the same as for the build. What changes is the number of hours, and the same factors drive it every time:

• User types and workflows. Every role, say customer, staff and admin, needs its screens and permissions worked out.
• Integrations. Each connection to an accounting tool such as Xero, a CRM such as HubSpot or a national eID such as BankID or MitID means reading documentation and getting test access.
• An existing system or old data. Anything that gets replaced or migrated has to be understood first.
• Stakeholders. Five people who all need a say mean more meetings and more rounds than one person who can decide the same day.
• Design fidelity. Rough wireframes are quick. A clickable prototype or polished designs take far longer.
• Compliance. GDPR, data residency in the EU, accessibility or industry rules have to be pinned down before anyone can estimate them.

Who runs the discovery matters too. A senior who has built similar products tends to know where the problems hide and spends fewer hours finding them. If you're weighing a freelancer against an agency or an offshore team, my developer cost comparison covers what each option really costs.

When a discovery phase is a waste of money

Discovery isn't always the right investment. These are the situations where I'd usually advise against it:

• The job is small and well defined. An integration or change under roughly 100 hours can usually be estimated straight from a good conversation and a written brief.
• You already have the clarity. If you hold a solid spec with wireframes and technical requirements, it may only need a review and an estimate, not a new discovery.
• You don't know whether anyone will pay. Then it's too early for discovery and for development. Talk to potential customers, or test a landing page or a clickable prototype first. My breakdown of MVP development cost covers when a first version is worth building.
• Off-the-shelf software already does the job. If a SaaS tool at a monthly fee covers your needs, try it first. A good discovery can end with exactly that recommendation.

Next steps: get ready for discovery

The more you prepare before kickoff, the fewer paid hours go into digging things up. Run through this list before you book a discovery phase, with me or anyone else.

How my fixed-price discovery phase works, and how a fixed price for the build follows from it, is explained on my pricing page. You'll deal with me directly, not a sales rep, and you'll hear back within one business day.

Frequently asked questions

Should the discovery fee be credited against the build?+

It depends on the supplier, and neither approach is wrong. Some credit all or part of the fee if you go ahead with them. Others don't, because the discovery output is a standalone deliverable you can take anywhere. What matters is that the agreement says which it is. Be wary of free discovery: the cost is usually recovered somewhere else, often by leaving you feeling tied to that supplier.

Can I take the discovery output to another developer or agency?+

You should be able to, and the agreement should say so explicitly. A useful discovery is written so that any competent developer can price and build from it. If the supplier keeps the rights or forbids sharing, you've paid for a sales document rather than a plan. Ask for the material in formats you can open without the supplier's tools.

Is a discovery phase the same as a discovery workshop?+

Not quite. A workshop is usually one part of a discovery phase, often the kickoff. The phase also covers the work between meetings: reviewing existing systems, checking integrations, sketching screens and building the estimate. A one-day workshop on its own rarely produces numbers you can budget against, so check what happens after it before you compare prices.

What if discovery shows my budget is too small?+

Then discovery has done its job. With a prioritized feature list you can cut a first version down to fit the budget, usually by pushing integrations, advanced roles or automation to a later phase. The alternative is finding out the same thing halfway through the build, with half the money gone. Sometimes the honest answer is to wait, or to start with off-the-shelf software.