From Lab to Market: A Founder's Five-Phase Playbook
What each Shipyards phase demands, the documents you need before you ask for money, and the three failure modes that end most technical startups.
Nia PattersonHead of Shipyards
18 min read
Technical founders rarely fail because the technology did not work. They fail because they built something nobody was waiting for, founded on intellectual property they did not own, or raised against a story rather than a system. Shipyards is five phases designed around exactly those three endings.
Phase one — Pioneer: deciding whether to start at all
Pioneer is ten classes and it is deliberately the easiest phase to leave. Its job is not to encourage you; it is to help you make an informed decision about the next two years of your life.
By the end of Pioneer a founder has three artefacts:
- A written problem statement. One page, naming a specific group of people, a specific situation, and what it currently costs them. If you cannot name the person, you do not have a problem statement — you have a technology looking for one.
- A customer hypothesis. Who pays, how much, and why they would switch from whatever they do now, including doing nothing, which is your most common competitor.
- A founder self-assessment. What can you build alone, what must you hire or partner for, and how long can you personally sustain no income?
Roughly a third of Pioneer participants decide not to continue, and we treat that as the phase working. A decision not to found, made in ten weeks with good information, is a vastly better outcome than the same decision made in two years with worse.
Phase two — Inceptor: the pre-filtration system
Inceptor is where the uncomfortable conversations happen. It advises prospective startups on processes, needs and documentation to minimise the difficulties and mistakes common in the startup industry — and it is the phase most likely to end with a no.
Three workstreams run in parallel over roughly twelve weeks.
| Workstream | What you produce | The gate |
|---|---|---|
| Market validation | Twenty structured customer conversations, written up | Evidence of a problem people already spend money or time on |
| Entity & IP | Incorporation, founder agreements, IP assignment, cap table | Clean ownership with no unresolved prior claims |
| Technical feasibility | A narrow prototype of the riskiest component | The hardest part demonstrably works, or is bounded |
The IP workstream deserves particular attention because it is the one founders most want to defer and the one that most often becomes fatal. Work done during employment, in a university laboratory, or with an unpaid collaborator has an owner, and it may not be you. Resolving that costs a few thousand dollars in Inceptor. Discovering it during diligence costs the round.
How to run the twenty conversations
Founders consistently run these badly on the first attempt, because the instinct is to pitch. The purpose is the opposite: to find out whether the problem you imagine is the problem they have.
- Do not describe your solution until the last five minutes. Once you do, the person becomes polite and the data becomes worthless.
- Ask about the last time, not about generally. 'Tell me about the last time this happened' produces facts. 'Would you use a tool that...' produces courtesy.
- Ask what they currently spend. Money, hours, or headcount. A problem with no current expenditure attached is a problem people have learned to live with.
- Ask who else has this problem and would talk to you. A referral is a stronger signal than anything they say about your idea.
- Write it up within an hour. Structured notes, same template each time, so that patterns across twenty conversations become visible rather than remembered.
At twenty conversations you will either have a clear, repeated pattern — the same problem described in the same terms by people who do not know each other — or you will have twenty different problems. The second outcome is not a failure of the interviews; it is the answer.
Phase three — Incubator: building the thing
With our incubator partnerships we help secure resources, funding, advisory and experienced mentor guidance to build out products in depth. This is the longest phase and the one where technical founders are most comfortable — which is precisely the risk.
Three disciplines we enforce, because comfort is the enemy here:
- Ship to real users every two weeks, without exception. Not a demo, not a staging environment. Real people, real data, real consequences.
- Instrument before you build. If you cannot say what would count as this feature working, you are not ready to build it. This single rule removes a remarkable amount of wasted effort.
- Write down what you expected before each release. Then compare. Founders who do this develop calibrated intuition within months; founders who do not remain confidently wrong for years.
The most dangerous phase for a technical founder is the one where they get to do the thing they are good at. Comfort is not progress.
Nia Patterson, Head of Shipyards, BTMP
Phase four — Accelerator: the last mile to market
Our accelerator platforms provide startups with everything needed to finish the last mile to market. In practice that means four things that technical founders systematically under-invest in.
| Area | The question it answers | Deliverable |
|---|---|---|
| Pricing | What is this worth to the customer, not what did it cost you | A tested price with rationale |
| Positioning | What category are you in, and who do you displace | A one-sentence position, tested on strangers |
| Compliance | What will block an enterprise purchase | A completed security and data questionnaire |
| Hiring | Who are the first three, and what do they own | Written role definitions and a hiring plan |
| Investor materials | Is the loop instrumented and honest | A data room that survives diligence |
On investor materials specifically: we do not let a founder open a data room until the acquisition-activation-retention-revenue loop is instrumented and the numbers are reproducible from raw data. Investors fund evidence of a working system. A deck that describes a system without the instrumentation behind it is a story, and stories get funded exactly once, at a price the founder later regrets.
Phase five — Post Modern Mentorship: after it starts working
Even after going to market, startups often need extra support on marketing, advertising and handling growth and scale. Phase five exists because the support infrastructure around founders evaporates at precisely the moment the problems change character.
What founders bring to this phase, in rough order of frequency:
- The first difficult firing, usually of someone who was excellent at ten people and is not at forty.
- A co-founder relationship that worked under pressure and is failing under success.
- The realisation that the thing they are now doing all day is not the thing they wanted to build a company to do.
- Board dynamics, which nobody prepares first-time founders for.
- Whether to raise again, and what it costs in control.
Access is open indefinitely to any Shipyards alumnus. It is the least structured phase and the one alumni use most.
The three endings, and which phase catches each
To close the loop on where we started. These are the three failure modes the programme is built around, and the gate that is designed to catch each.
| Failure mode | How it presents | Caught by |
|---|---|---|
| Building before validating | Eighteen months of development, no customers | Inceptor market validation gate |
| Unresolved IP ownership | A claim surfaces during diligence | Inceptor entity & IP gate |
| Raising against a story | A round closes, then the metrics do not reproduce | Accelerator investor-readiness review |
None of these gates is pleasant. All three are cheaper than the alternative, and the founders who tell us Shipyards was most valuable are disproportionately the ones we told 'not this' rather than 'yes, go'.
Frequently asked questions
Pioneer takes none. Later phases involve partner resources whose terms are disclosed in full before you enter the phase, and no founder is ever asked to commit before seeing them.
Yes, if you have the evidence that phase's entry gate requires. Say so in your application and attach the artefacts — a problem statement, customer write-ups, or a working product.
Many of our founders start solo. Pioneer's self-assessment is partly designed to clarify what you genuinely need a partner for, versus what you can hire, contract or defer.
Pioneer is around four hours a week. Inceptor is closer to fifteen. Incubator and Accelerator assume the company is your primary occupation.
Key takeaways
- Pioneer's job is an informed decision, not encouragement — a third of participants rightly stop.
- Resolve IP ownership in Inceptor, not in diligence; the price difference is the whole round.
- Run twenty customer conversations about the last time, not about generally.
- Instrument before you build: if you cannot define working, you are not ready to start.
- Do not open a data room until the acquisition-to-revenue loop reproduces from raw data.
Nia Patterson
Head of Shipyards, BTMP
Writes here about the parts of the work that are rarely taught and always tested. Mentors on the programme and reviews cohort capstones.
