A Lean LaunchPad Course Outline, Updated for Students Who Arrive With a Product
Steve Blank built the Lean LaunchPad at Stanford in 2011. It became the structure underneath NSF I-Corps and it is probably the most copied entrepreneurship course design in existence, for a good reason: the weekly cadence of getting out of the building and reporting back is the only part of entrepreneurship education that reliably transfers.
Blank himself has been direct about what AI changed. In his account of his Spring 2026 Stanford class, every presenting team arrived with a finished, functional product built before the course started. His conclusion is narrower than the headlines suggested: Customer Development, the canvas, hypotheses, interviews, and pivots all survive. What broke is the MVP as proof. Building something stopped being evidence that a team learned anything, because building something now costs a weekend.
That is a one-component failure in an otherwise working machine. Here is the outline with that component replaced.
What stays, unchanged
- Weekly customer contact. The cadence is the method. A course that collects interviews twice a term is referencing Lean LaunchPad, not running it.
- The canvas as the working document, revised weekly, not written once.
- Class time spent critiquing evidence, not ideas. The instructor’s job is to ask “how do you know that” until it stops being answerable.
- Teams, not individuals. Customer discovery at volume needs more than one person and the arguing is productive.
What changes
The build requirement moves from “ship an MVP” to “run a test that costs someone something.”
That is the substitution. An MVP used to be expensive, which is what made it meaningful. Now the scarce and meaningful thing is getting a stranger to spend money, time, or inconvenience. A signed letter of intent, a deposit, a scheduled pilot, a person who switched tools for a week: those cost the customer something, and no model can generate one.
The ten-week outline
Week 1. Business model hypotheses. Teams fill the canvas in a single session and immediately mark every block as assumed. The speed is the lesson: nine boxes, forty minutes, and nothing in them is known.
Week 2. Value proposition and customer segment. Narrow until the segment is findable. The test is whether the team can name where to locate fifty of these people this week.
Out of the building: five interviews, segment validation only.
Week 3. Customer discovery mechanics and first results. Teams report. The instructor’s only job this week is to separate what someone said from what the team wished they said. Do this ruthlessly once and it sets the tone for the term.
Out of the building: ten interviews.
Week 4. Channels. The most neglected block and a frequent cause of death. How does this person find out the thing exists, and who already has their attention?
Out of the building: five channel conversations, which are a different kind of interview and worth saying so.
Week 5. Revenue and pricing. Not spreadsheets. What do they pay now, to whom, and what would switching cost them beyond money?
Out of the building: ten interviews including explicit price conversations.
Week 6. Pivot or persevere, formally. Published decision point. Narrowing the segment counts and is the most common right answer. Without a scheduled window, teams defend dead ideas because the sunk cost is a grade.
Week 7. Partners, resources, activities. What has to be true operationally, and which of those is the team assuming rather than knowing?
Out of the building: five interviews with potential partners or suppliers.
Week 8. Costs, and the test that costs someone something. Teams design the commitment test described above and run it.
Out of the building: attempt to secure one real commitment.
Week 9. Evidence assembly. No new interviews. Teams build the file: what we believed in week one, what we tested, what we learned, what moved.
Week 10. Lessons learned presentations. Blank’s original format and still the right one. Teams present what changed, not what they built. Outside questioners if you can get them.
The weekly rhythm that makes it work
Every session runs the same way and the sameness is a feature:
- Ten minutes per team reporting what they learned since last week
- Instructor and peers ask “how do you know that”
- New material for the week, kept short
- Teams state what they will test before the next session
Most of the teaching happens in step two. A good Lean LaunchPad instructor talks less than students expect and asks the same question more often than students enjoy.
The grading problem this creates, and the fix
Weekly cadence means weekly evidence, which means an instructor reading a large volume of interview notes. This is the practical reason Lean LaunchPad courses drift back toward milestone grading and lose their effectiveness.
Two things help. First, grade the delta, not the volume: what changed on the canvas and why. Second, require interviews to be logged with dates, names, roles, and sourcing as they happen, so the record exists continuously instead of being assembled retrospectively. Retrospective assembly is both a grading problem and the point at which fabrication becomes attractive.
On interview counts
Blank’s Stanford course and NSF I-Corps both push teams toward roughly 100 interviews. That is right for full-time intensive programs, and unrealistic for undergraduates taking four other courses.
For a semester course, ten to twenty-five per team is honest. Past about ten, sourcing quality matters much more than count. Twenty interviews with the team’s own network teaches less than eight with strangers, and the rubric should say so.
Try the ToolSuite free with your class
Customer discovery and business model canvas tools built for coursework. Free to start, with academic pricing for programs.
Check out our software free →
The weekly-cadence problem is visibility: knowing which team is in trouble on Tuesday rather than at the midterm. The customer discovery and canvas tools in Startup Ignition ToolSuite log interviews with dates and sourcing as students run them, and the canvas keeps version history so the weekly delta is visible without a separate submission. Academic pricing is available for accredited programs.
If you want the detail of Blank’s argument about what AI did to the course, I summarised the series here. The grading side is in this customer discovery rubric.