For founders
"What do I get each week?" The MVP sprint, week by week
Three weeks from an agreed scope to a deployed product. A prototype on the first Friday, a staging link you can use on the second, a live product on the third. Here is what lands on your desk each week, and what I need from you to keep it there.
Prashant Kumar Singh · · 4 min read
In short
- Week 1 ends with a clickable prototype and the data model agreed. Week 2 ends with a staging link where the core flow works end to end. Week 3 ends with the product live, edge cases handled, and handover documentation.
- Every Friday: a written update with done, next, risks and decisions needed, whether the news is good or not.
- The scope is one flow, fixed before day one in the free teardown, which is what keeps three weeks at three weeks.
Founders who have been through a bad outsourcing project ask one question before anything else: what do I actually get, and when? The answer for the $3,000 MVP sprint is specific enough to write down, so here it is, week by week, with what you see on each Friday and what I need from you to keep the dates.
Before day one
The free teardown has already happened. It ended with a one-page scope: the one flow, the screens, the stack, what is explicitly out. The price is fixed against that page. If you did a design sprint first, the prototype and scope from it are the starting point and week one is shorter.
Week 1: design
Monday and Tuesday: the flow on paper, then screens in Figma, every state, real copy. Wednesday: the data model, what gets stored, how it relates, which parts of the flow read and write it, agreed in writing. Thursday: the screens linked into a clickable prototype. Friday: you use the prototype, tell me where you got stuck, and the first written update arrives.
What you get on Friday: a clickable prototype of the whole flow and the agreed data model. What I need from you: an hour using the prototype and your notes by Monday.
Week 2: build
The core flow is built end to end: auth, the screens, the backend, the database, the one integration the flow needs if there is one. If the product has an AI feature, this is when it is wired in, with sources, a fallback and a cost cap. By Friday the flow works on a staging URL you can log into.
What you get on Friday: a staging link where you can complete the flow with real data. What I need from you: use it as if you were a customer, and send the three things that bothered you most.
Week 3: ship
Edge cases from your week-two notes and from my own testing. Error states, empty states, the slow-connection case, the wrong-input case. Then deployment to your cloud account, with the domain, monitoring, and a handover document that explains how to run it, where the keys are, and what to do when something breaks. Friday: live.
What you get on Friday: the product live on your domain, the repository, the design files, the handover document, and 30 days of support starting. What I need from you: access to the cloud account and the domain by Wednesday.
| Friday | Delivered | You do |
|---|---|---|
| Week 1 | Clickable prototype, data model agreed | Use it, send notes by Monday |
| Week 2 | Staging link, core flow end to end | Use it as a customer, send three things |
| Week 3 | Live product, repo, designs, handover, support starts | Provide cloud and domain access by Wednesday |
What moves the dates, and what does not
Two things move the dates: a decision that waits more than a few days, and new scope. The first is why the Friday update has a decisions-needed section with a recommendation for each, so a one-word reply is enough. The second is why new ideas are written down and quoted rather than squeezed in. Everything else, including the normal surprises of building software, is inside the three weeks, because the scope was cut to fit them.
Questions people ask
- How long does it take to build an MVP?
- Three weeks for one complete flow: design and prototype in week one, the core flow built on staging in week two, edge cases and deployment in week three. The scope is fixed beforehand so the timeline holds.
- What do you deliver each week of an MVP sprint?
- Week one: a clickable prototype and the agreed data model. Week two: a staging link where the flow works end to end. Week three: the live product, repository, design files, handover documentation and 30 days of support.
- What can delay an MVP sprint?
- A client decision that waits more than a few days, or new scope added mid-sprint. Both are handled by the Friday update's decisions-needed section and by quoting new ideas rather than absorbing them.
If you're a founder
Send your product or idea. The teardown comes back with the one-page scope this sprint would run on, and a start date.
If you run an AI lab or data team
A ten-task pilot has the same rhythm: a report at the end of the week, decisions needed listed, and nothing added to the scope without being written down.
Sources