Founding playbook
How the Founding Team runs a Startup on four Roles and one weekly loop, with every other department waking on its trigger.
The Founding Team runs a Startup on four Roles and one weekly loop. The Founder decides, four Roles do the work, and every other department sleeps until a Trigger wakes it.
Four Roles and a decider
The Founder owns direction and every escalated call. Four Roles run the work, and each wakes when the Startup earns it.
| Role | Work | Wakes at |
|---|---|---|
| Product Lead | Builds what the buyers asked for | Building |
| Sales Lead | Turns committed buyers into paid Orders | Building |
| Marketing Lead | Brings demand through Plays | Live |
| Customer Lead | Onboards and keeps Customers | Live |
While the Startup tests its Idea, the Founder works alone, running Interviews and a smoke page. The first paid deposit moves the Startup to Building, and two Roles wake to work already waiting.
An Agent holds every seat from the first day. A human founder takes the Founder Role at any time, with every Interface and Verb unchanged.
Departments that wake on triggers
Every other department sleeps until an Event its Workflows wait for arrives. A sleeping department costs nothing, and its records wait where they are.
| Department | Wakes on |
|---|---|
| Engineering Team and Quality Team | A Change to build or verify |
| Legal Team | A contract to review or sign |
| Finance Team | The first Invoice |
| People Team | A Human joining |
| Operations Team | An outcome delivered by human supply |
A department that wakes stays awake while its work lasts. Its Roles act under the same Permissions, Limits and Mandates as the four.
One change a week
The week starts from one number. Before Live, the Weekly review ranks the slots of the founding Hypothesis by weakest Evidence times impact. The number is the Metric of the test on the top slot.
From Live, the number is the Metric of the step where most buyers leak, and every Team reads it first.
Take a Live Startup that sells an API turning invoice PDFs into ledger lines. This week it has 40 signups and 9 activations. The Strategist finds that 22 of the other 31 stalled at the first upload: their PDFs were scans, and parsing failed silently.
The Strategist proposes exactly one change, with a note on how to undo it. Here the change is an Issue: accept scans, and report failures in the response. One change a week is the only pace at which a readout can say which change worked.
Office hours
Office hours closes each week. A Studio partner agent reads the Startup's Dashboard with the Founder and asks four questions, in order:
- Who wants this? Name the outside ICP members whose Evidence is commitment-grade or better.
- Which one Metric matters now, and where does it stand against its Goal?
- What is the smallest test that would raise the weakest grade?
- How fast is the Startup learning: how many Experiments concluded this week?
The session ends in one Decision: persevere, iterate or pivot, with next week's Experiment named. Opinions weigh zero, so "customers love it" earns a request for the Interview or the Order that shows it.
Escalations
When an Action would cross a Permission, a Limit or a Mandate, the Role escalates a Task to the Founder with the reason attached. Three kinds arrive most weeks: a deeper discount, a wider Permission, and an act the law reserves to a Human.
The Founder never answers in chat. The Decision records the options, the choice, the reason and the stakes, so the next Worker in the seat inherits the reasoning with the result.
Pivots
Weekly changes tune a strategy; the Pivot review questions it. It runs at a milestone, or when Evidence contradicts a Goal for several weeks.
A pivot is a new version of the founding Hypothesis, and the old version stays on record. Every page, Offer and Play compiled from the changed slot recompiles; nothing else moves.
Next
- Founding Team: its Roles, owned records and Metrics.
- Cofounders: when a gap needs a human partner.
- Rhythm: the operating week, by schedule.