What a discovery sprint should give you in two to four weeks

By Martin Chadwick, founder · 25 September 2026 · 2 min read

In short

A discovery sprint is two to four weeks of research, workshops and technical investigation before you commit to a build.
You should leave with evidence about users and the operation, a scoped first release, a prototype you can test and a costed roadmap.
It works best with a named decision-maker, access to real users and real data, and a willingness to change the brief.
The output should be useful with or without the team that ran it.
Discovery is also what makes funding applications and investor decks credible.

Discovery has a reputation problem. Done badly, it is a few weeks of workshops that end in a slide deck and an invoice. Done well, it is the cheapest way to avoid building the wrong thing.

Here is what a good discovery sprint should give you.

Week one: understand the operation

Stakeholder workshops to agree the goal, the constraints and how success will be measured.
User research with the people who will use the product. Interviews and observation, not just surveys. AI can speed up transcripts and tagging, but AI tools for UX cannot stand in for real users.
The current state, mapped: tools, data, handovers and the workarounds people have built. If a product already exists, a UX audit is a quick way to start.

Week two: shape the answer

Journeys for the most important tasks, showing where the product will change them.
Benchmarks of how others solve similar problems.
Technical discovery: data sources, integrations, hosting and the risks that could sink the build.

Weeks three and four: prove it

A clickable prototype of the core journey, tested with real users.
A scoped first release: what is in, what is out and why.
A costed roadmap with phases, milestones and the key risks.

Shorter sprints compress these steps. Longer ones add more testing rounds.

What makes it work

A decision-maker in the room who can agree scope changes on the spot.
Access to real users and real data. Discovery on assumptions produces a plan built on assumptions.
Willingness to change the brief. The most valuable outcome is sometimes "build something smaller" or "do not build that yet".

Useful with or without us

The output should belong to you and stand on its own. A good discovery pack can go to an internal team, another supplier, a funder or an investor. If it only makes sense to the people who wrote it, it has failed.

Discovery and funding

Funders and investors want evidence, a clear scope and a credible plan. Discovery produces exactly that. A tested prototype and a costed first phase make a much stronger application. See our guide to funding a digital product in Greater Manchester.

Our discovery sprints run for two to four weeks at a fixed scope agreed up front. Start a conversation.

Common questions

What is a discovery sprint?

A short, fixed-length phase of research, workshops, technical investigation and prototyping that turns an idea or problem into a scoped, costed plan before a build begins.

How long should discovery take?

Two to four weeks for most products. Longer discovery tends to drift unless the problem is very large or heavily regulated.

What should we have at the end of discovery?

Research findings, a prioritised scope for the first release, a clickable prototype, a technical approach with risks, and a costed roadmap.

Martin Chadwick is the founder of Third Eye Open, a design and engineering studio at Campfield House, Manchester. He has spent over fifteen years designing and building digital products, and runs TEO’s own products, TEO Sense and ScoreUX.