Release and Metrics
Two artifacts, each doing what its medium is good at: a 3-minute trailer that proves the software runs and that someone might actually watch, and a 2-page report carrying the numbers. Splitting them is deliberate. A twelve-minute video is not something a recruiter, a future teammate, or your partner’s boss ever finishes, and measured results belong in text where they can be checked, quoted in a resume, and read again next year.
You have demoed every sprint for three terms. This is the polished cut of the one flow that matters.
The Trailer (3 minutes, hard cap)
Section titled “The Trailer (3 minutes, hard cap)”Upload to media.oregonstate.edu as unlisted; public platforms in addition are welcome. Closed captions or a transcript are required. The Presentations guide covers structure and delivery.
Three minutes forces one decision: which single thing do you show? Answer it before you record.
- The problem, in one sentence (about 20 seconds). Who has it and what it costs them. No preamble, no team introductions, no agenda slide.
- The software running (about 2 minutes 10). One complete flow end to end from a realistic entry point. Product and tool teams: a real user task. Research and pipeline teams: the pipeline running on real data and producing the result. Narrate what is happening. No silence longer than five seconds, and no slides where you could show the thing itself.
- What it achieved, in one line (about 20 seconds). One real number, and where the report is.
Say plainly if a feature shown is partial. Fabricated footage fails the assignment.
The Report (2 pages, PDF)
Section titled “The Report (2 pages, PDF)”This is the graded evidence and the durable artifact. Convert it from Markdown with the one-liner so the cover carries your project ID, name, and team.
- Problem and scope: the problem concretely, who experiences it, and what your team deliberately left out.
- Delivered: the features, findings, or merged contributions that exist, with what is fully working, partial, or descoped stated as such.
- Metrics: restate the 2 or 3 success metrics from
docs/shipped.md. Report a real measured value for at least one, with instrumentation evidence embedded: an analytics screenshot, benchmark output, install counts, user task results. Winter’s instrumentation commitment comes due here. - Delivered versus the Definition of Shipped: your category’s target rung, where you actually landed, and the gap with concrete numbers. Patches accepted, users reached, deployment status, findings produced.
- Limitations and next step: the one or two most important lessons and the highest-priority thing a successor should do.
- Links: the trailer, the repository or archive, the Landing Page, and the release tag. Test every link in a private browser window before submitting; a link graders cannot open scores zero on the affected criteria.
Rubric (100 points)
Section titled “Rubric (100 points)”| Criterion | Points | Outcome |
|---|---|---|
| Trailer: one complete flow shown end to end from a realistic entry point, within 3 minutes, captioned, honest about partial features | 20 | SO2 |
| Problem and scope: concrete, audience-appropriate, with the cost to the user and what was deliberately left out | 15 | SO3 |
Metrics evidence: metrics match docs/shipped.md, at least one reported with a real measured value and embedded instrumentation evidence |
30 | SO2 |
| Honest outcomes: delivered versus the Definition of Shipped with concrete numbers, lessons, and a prioritized next step | 25 | SO6 |
| Report quality: within 2 pages, structured, every link tested and working | 10 | SO3 |
AI use: AI may help with scripting, editing, and drafting the report. Fabricating demo footage, metrics, findings, or user feedback fails the assignment; a smaller true number always beats a bigger invented one. The outcomes section is graded on honesty, and “we did not reach the bar, here is exactly how far we got and why” scores well.
Activities That Prepare This
Section titled “Activities That Prepare This”Run an Acceptance Pass for the Honest outcomes criterion, and Learn from TED Presentation Advice for the Trailer criterion. Identify Success Metrics from winter is what makes the metrics-evidence criterion answerable.
By project category:
| Category | Activities |
|---|---|
| FOSS | Software Release and Versioning: your release evidence is upstream’s, so report merges and releases in their terms |
| Research | Make Your Artifact Reproducible and Keep an Experiment Log, which together are what turns a result into reportable evidence |
| Consultancy | External Feedback Session with the partner, since your metric is largely their adoption |
| New Product or Game | Identify Success Metrics, instrumented in winter and reported now |
Browse communication and user when these run out.