11. Presentation Winning
A good hackathon pitch is not an essay. It is a compressed story that helps the judge understand value quickly.
Winning pitch structure
flowchart TD
A[Problem] --> B[User] --> C[Solution] --> D[Demo] --> E[Impact] --> F[Future]
Core pitch template
- Problem
- User
- Current pain
- Your solution
- Live demo
- Why it matters
- What is next
Judge psychology in the room
Judges respond well to:
- clarity,
- confidence,
- visible progress,
- and believable usefulness.
They respond poorly to:
- overexplaining,
- vague AI claims,
- broken demos,
- and giant scopes.
Demo psychology
Your demo should feel:
- smooth,
- short,
- live,
- and easy to follow.
Recovery plan if the demo breaks
- Stay calm.
- Show screenshots.
- Show the deployed URL.
- Explain the expected behavior.
- Keep the pitch moving.
PPT psychology
| Slide type |
Goal |
| Title |
Make the project memorable |
| Problem |
Create urgency |
| Solution |
Show clarity |
| Demo |
Prove it works |
| Impact |
Show why it matters |
| Future |
Show ambition without overpromising |
Copy-paste pitch template
30-second version
“We built a tool for [user] who struggles with [problem]. It helps by [solution]. Here is the live workflow, and this is why it saves time or reduces friction.”
60-second version
“[User] currently deals with [problem]. Existing solutions are hard because [reason]. We built [name], which [core action]. In the demo, you will see [flow]. This matters because [impact].”
Slide design rules
- One idea per slide
- Use large text
- Avoid walls of text
- Show real screenshots
- Keep the flow simple
- Highlight the result, not the implementation first
Practice checklist
Best rule
The pitch should make the judge think, “This is clear, useful, and finished.”
The 3-minute structure judges can follow
| Time |
Say |
Show |
Avoid |
| 0:00–0:25 |
One-sentence problem + who feels it |
Problem slide, one stat or quote |
Origin story, team intros |
| 0:25–1:00 |
Your solution in one line + the core loop |
Live product, golden path only |
Feature list, architecture |
| 1:00–1:50 |
Live demo, narrated click-by-click |
Real data, seeded account |
Typing, sign-up flows, logs |
| 1:50–2:25 |
Why it matters + what you validated |
Before/after, tiny metric |
“AI-powered” without proof |
| 2:25–3:00 |
What’s next + ask |
Roadmap (3 items max), live link + QR |
Overpromising, ending on “questions?” |
Script skeleton (fill the brackets, then cut 20%):
“[User] wastes [time/money] on [problem] because [current workaround] is
[slow/expensive/confusing]. We built [name]: [one-line solution].
Watch — [click 1] → [click 2] → [result]. That [saves X / removes Y].
Next we’d [1, 2, 3]. Try it now at [URL].”
Demo scripting rules
- Golden path only. One user, one task, under 90 seconds.
- Seed everything. Logged-in account, realistic data, no empty states.
- Narrate clicks before you click. “I’ll upload a bill — watch the total extract.”
- No live typing. Pre-fill inputs; paste long text.
- No sign-up on stage. Auth is the #1 demo killer.
- Have a
?demo=1 mode. One click loads sample data if an API is down.
- Rehearse the fail line: “The live API is rate-limited right now — here’s
the recording from 20 minutes ago showing the same flow.”
Booth vs stage
- Booth (science-fair): judges come to you for 3–5 min. Open with the
product, not slides. Let them click within 30 seconds. Keep a one-pager
with the link + QR on the table.
- Stage (timed pitch): slides + strict clock. Memorize the first and last
20 seconds; those decide recall. End 10 seconds early — running over reads
as unfinished.
Q&A bank — the 15 questions judges actually ask
- Who is this for, specifically?
- What problem does it solve that existing tools don’t?
- How did you validate anyone wants this?
- What did you build vs what is mocked?
- What breaks if 100 users try this now?
- How does the AI part work — model, prompt, fallback?
- What happens when the API is down or rate-limited?
- Where is data stored and who can see it?
- Why this stack — what did it let you ship faster?
- What would you cut if you had 4 fewer hours?
- What would you build with 4 more weeks?
- How would you get your first 100 users?
- What did each teammate own?
- What was the hardest bug and how did you fix it?
- Can I try it myself right now? (Always answerable with URL + demo login.)
Answer formula: direct answer (10 s) → evidence (20 s) → scope honesty (10 s).
Never bluff a live capability — offer the recording instead.
Body language + delivery
- One speaker drives; others handle clicks and Q&A. Switching speakers mid-demo wastes time.
- Stand where you can see the screen without turning your back.
- Speak 10% slower than feels natural; hackathon rooms are noisy.
- Point at the result, not the code, unless asked for depth.
- End on the link: say it, show the QR, and stop. Don’t trail off.
Timing drill (do this twice)
- Run
python3 tools/pitch-timer.py --file pitch.txt — targets:
30 s ≈ 75–90 words, 60 s ≈ 150–180, 2 min ≈ 300–360, 3 min ≈ 450–540.
- Record one full run on your phone. Cut filler until you’re 10 s under.
- Do one run with wifi off (backup recording + screenshots only).
templates/pitch-script.md — 30 s / 2 min / 5 min skeletons.
examples/example-pitch.md — worked example pitch.
tools/pitch-timer.py — word-count + pacing check.
- Sections
20-judging-insider/, 29-storytelling/ — scoring and narrative depth.
Last verified: 2026-09-10.