Melo: a Product Management squad turned an undefined problem into a live MVP.
How a squad of Product Management residents took an undefined problem, investigated it with real people, and shipped a working MVP for a pre-seed startup.
The engagement
Client
Forklore Labs, a fictional client built for the Residency: a pre-seed, pre-product startup with a solo founder, £150K raised, nine months of runway and one contract engineer on standby.
Mandate
Investigate how households decide, plan and shop for meals, then recommend: build, reshape the idea, or don’t build.
The team
A squad of Product Management residents working remotely, with a nominated Team Lead and Assistant Team Lead.
Mentors
Abisola Rachael (Weeks 1–4, Product Management Instructor, Heala Tech) and Segun Sobodu (Weeks 5–8, Product Manager, Opera Mini).
Cadence
Thursday working sessions, Saturday sprint reviews, weekly 1:1 mentor review, four decision gates and three surprise pressure events.
Output
An evidence-backed recommendation, a full set of product documents, and Melo, a live MVP now open for early-access testing.
An ambiguous problem, a limited budget, and a founder who needed a decision.
The founder had spent eight years in grocery retail and had watched shoppers buy the same trolley every week and throw much of it away. They were sure there was a business somewhere in how busy people decide what to eat, budget, shop and waste food. But they had deliberately not defined a product.
The brief was simple: find out who really feels this problem, what the real pain is, why current tools fail them, and whether any of it is worth building. If the honest answer was “don’t build”, that would save the runway. Any team whose first-week documents already assumed an app was sent back to the evidence.
That is what makes this engagement realistic. Real product managers rarely get a tidy feature request. They get an ambiguous problem, a limited budget and a stakeholder who needs a decision.
Brief to shipped MVP, week by week.
Onboarding and discovery
Team charter and board, stakeholder map and decision rights (DACI), risk log, founder kickoff Q&A, hypothesis map, research plan and a neutral interview guide.
ResearchGate A · evidence sufficiency
12+ user interviews, 40+ survey responses and diary studies; synthesis, personas, empathy maps, jobs-to-be-done, journey map and a competitor scan.
Market and strategyGate B · problem selection
TAM/SAM/SOM sizing, PESTLE and SWOT, an Opportunity Brief and solution tree, a vision and strategy one-pager, and concepts scored with RICE.
Definition
Story map, prioritised backlog, roadmap and OKRs, a one-page BRD, PRD v1, user stories with acceptance criteria, and agreed Definitions of Ready and Done.
Build
Success-metrics framework, event-tracking spec, experiment plan, Figma prototype v1 and a feasibility session with the engineer.
Validate and price
Five validation sessions, prototype v2, pricing and willingness-to-pay work, go-to-market plan and positioning, launch risk review.
Stakeholder reviewGate C
Stakeholder update deck for the founder, assumption validation notes, revised documents (v2) and a retrospective.
ShipGate D
Product case study, recommendation memo, live MVP and a final panel defence with a demo.
Three surprise events, no warning.
A new competitor launches
Mid-way through Week 3, the founder forwarded news of a new competitor and asked, “Does this kill us?” The team had 48 hours to reply with a one-page, evidence-based note.
The engineer pushes back
In Week 5 the contract engineer (played by a mentor) challenged the most expensive assumption in the concept. The team had to cut scope without bending the truth.
Every claim needs a source
Assumptions stayed labelled as assumptions until evidence supported them. No interviews, quotes or survey answers could be invented, and AI drafts had to be understood and defended.
The plan was made on a calm day, and then real life broke it.
The research moved the team away from the obvious idea, which was yet another meal-planning or recipe app. Plenty of those already exist. People got home late, were too tired to cook, were missing one ingredient, had less money than expected, or opened the fridge and drew a blank. At that moment they had two bad choices: force the original plan or give up and wing it.
Melo was designed as the third option: a meal decision companion that starts with what changed and recommends one practical next step.
Melo asks three things, then points you down one of four paths.
What you planned, what changed, and what you have, then it recommends one next step and explains why, instead of handing you a long list of recipes.
Rescue
Use what you already have in the kitchen.
Adapt
Keep the idea of the planned meal but make it easier.
Top-up
Buy the one missing ingredient instead of abandoning the plan.
Order
Choose convenience without guilt when cooking doesn’t fit the day.
What the team chose not to promise.
Good product managers say what a product will not do. The team was explicit about Melo’s limits, and that scope discipline is visible on the live site.
- Direct ordering, delivery integration, and live restaurant or price data are not part of the first version.
- Melo does not promise the cheapest meal, and it is not medical nutrition advice or an allergen guarantee.
- The launch is an early-access trial, so real users shape what gets built next, and every recommendation asks for feedback.
Skills each resident can now walk an interviewer through.
Tools used hands-on included Jira, Notion, Figma, Miro or FigJam, Google Forms, and Amplitude or Mixpanel, with the live MVP hosted on Netlify. The project framework records 37 of 46 Product Management curriculum topics as applied hands-on during the Residency.
From Week 3, each resident also worked on getting hired.
- Three target job descriptions and a CV evidence log, leading to an ATS-checked CV (75%+ match) and a refreshed LinkedIn profile.
- Five to eight STARL interview stories built from their own Melo work, rehearsed through 100 practice questions.
- A group mock interview, an AI interview and an external examiner interview.
- A structured application plan (45 applications a week from Week 6) and three months of outcome support after the programme.
- A public Open Day where residents present the team project and their individual work to external judges.
Most courses leave you with a certificate. This one leaves you with a live product.
Residents finish with a live product, a full set of real documents, and specific stories about decisions they made under pressure. That is what hiring managers ask about.