What should an entrepreneurship capstone require students to produce?

An evidence file and a recommendation, not a launched company. Requiring a launch rewards the team with the easiest idea and punishes the team that correctly killed a hard one. The capstone deliverable that holds up is a documented record of what the team believed, what they tested, what they learned, and what they now recommend, with the recommendation permitted to be do not proceed.

How do you grade an entrepreneurship capstone fairly across teams?

Grade the delta rather than the outcome. Teams get wildly unequal ideas, markets, and luck, so scoring the result scores the draw. Scoring how much a team's understanding moved against a memo they wrote in week two, and how much evidence drove the movement, compares teams on the only thing they control.

How do you stop one student doing all the work in a capstone team?

Make the individual contribution visible in the artifact rather than in a peer survey at the end. Require interviews to be logged under the name of the person who ran them, and require each team member to personally run a minimum number. A team of four where one student logged eleven interviews and three logged none is not a peer-review problem, it is visible in the data before week six.

Should capstone teams pivot?

Yes, and the syllabus should make it cheap to do so before a fixed date. Teams that cannot pivot defend a dead idea for a semester because the sunk cost is a grade. A published pivot window, with a lighter deliverable attached, converts that from a failure into an expected step.

An Entrepreneurship Capstone That Produces Evidence, Not Demo Day

John Richards · November 2nd, 2026

The capstone is the course where entrepreneurship programs most often get graded by their own alumni, usually years later and usually harshly. The common complaint is not that it was too hard. It is that it was theater: a semester of preparation for a demo day that rewarded presentation and taught very little about whether anything was true.

Here is a shape that avoids that. It assumes a fifteen-week semester, teams of three to five, and students who have already had an introductory course.

What the capstone produces

Not a launched company. Not a demo day. An evidence file and a recommendation.

Requiring a launch sounds rigorous and is not. It rewards the team that picked the easiest idea and punishes the team that picked a hard one and correctly killed it. Those are backwards incentives for a course that is supposed to teach judgment.

The recommendation is permitted to be “do not proceed.” Say this in week one, put it in the syllabus, and repeat it at every milestone, because students will not believe it until they have seen a team do it and receive a good grade.

The six milestones

Milestone	Week	Weight
Team charter and belief memo	2	10%
Riskiest assumption test plan	4	10%
Evidence review one	6	15%
Pivot or persevere decision	9	15%
Evidence review two	12	25%
Final recommendation and defence	15	25%

Week 2: team charter and belief memo

Two documents. The charter covers who does what, how decisions get made, and what happens if someone stops contributing. It feels like busywork in week two and prevents the most common capstone failure in week ten.

The belief memo is the one that matters for grading: what the team believes about the customer, the problem, and willingness to pay, stated so it could be proven wrong. Fixed date, no revisions. Everything downstream is scored against it.

Week 4: riskiest assumption test plan

Which belief, if false, kills this? How will you test it, with whom, by when, and what result would change your mind?

Require a number. “We will interview restaurant owners” is not a plan. “We will interview twelve independent restaurant owners who do their own scheduling, by October 10, and if fewer than four describe this as a top-three weekly frustration we will drop the segment” is.

Week 6: evidence review one

First substantive checkpoint. Interviews logged with dates, names, roles, and sourcing. What contradicted you?

Grade this one firmly. Teams that coast through week six do not recover by week twelve, and a gentle grade here is a kindness that costs them the semester.

Week 9: pivot or persevere

A formal, published decision point. Teams present what they have learned and state whether they are continuing, narrowing, or changing.

The pivot window exists because without it, teams defend dead ideas for a full semester. The sunk cost is not emotional for a student, it is a grade, and you have to remove that penalty structurally or they will rationally refuse to change.

Narrowing the customer segment counts as a pivot and is the most common correct answer.

Week 12: evidence review two

The heaviest intermediate weight. The question is no longer “did you talk to people” but “what do you now know that you did not in week two, and what is it based on?”

Require a side-by-side with the week two memo. The comparison is the assignment.

Week 15: recommendation and defence

A written recommendation and a live defence, with questions from outside the teaching team. Invite an investor or operator. Students perform differently when the questioner has written cheques, and the questions are better.

No slide-deck grade. Teams may use slides. The grade attaches to the evidence and the reasoning.

The grading split that resists polish

Notice that presentation appears nowhere. That is deliberate, and it needs saying out loud more than once, because students arrive believing the capstone grade is a presentation grade and will allocate their last three weeks accordingly.

Within each milestone, I weight roughly:

Evidence quality and traceability: 40%
Disconfirming evidence reported: 25%
Reasoning and decision quality: 25%
Clarity of communication: 10%

Clarity earns a little because a recommendation nobody can follow is not useful. It earns only a little because AI now supplies polish on demand, and a rubric that pays for polish is paying for a commodity.

The free-rider problem, solved structurally

Every capstone instructor knows the team where one student does everything. Peer evaluations at the end of term catch it too late and turn into a social problem.

Fix it in the artifact. Interviews are logged under the name of the person who ran them, and every team member must personally run a minimum, say six across the semester. Now the distribution is visible in week six, not week fifteen, and the conversation you have is about data rather than about feelings.

This single change has done more for team dynamics in the cohorts I have run than any amount of charter language.

Adapting it for a quarter

Ten weeks: drop evidence review one and move the pivot window to week six. Keep both the belief memo and the final defence. The pivot window is the piece most often cut and the piece least able to be cut, because it is the only thing that lets a team recover from a bad week-two idea.

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 capstone-specific problem is visibility: knowing in week six which team is in trouble, without reading forty pages per team. The customer discovery tools in Startup Ignition ToolSuite log interviews per student with dates and sources, and the canvas keeps version history, so the delta the capstone is supposed to measure is visible without a separate submission. Academic pricing is available for accredited programs.

Back to All News