At a hackathon, the developers’ festival, or a startup pitching competition, the most tense moment is not when coding is finished. It is the live demo: putting the app you have built on screen and running it in front of hundreds of audience members and judges.
Salesforce’s Dreamforce Hackathon, held in San Francisco in 2025, explicitly included demo delivery itself as one of its four official judging criteria. However innovative the code, a team cannot receive a good assessment unless it clearly shows how the app works for a user on stage. As the equipment and output guidelines in the “Case Article Review” file also repeatedly emphasise, live demonstrations on stage are among the most difficult technical variables for event planners too. A clear, cool-headed back-end rule set is needed to run a demo stage fairly and smoothly amid uncontrollable variables such as Wi-Fi disconnections, laptop resolution conflicts and app bugs.
1. Decoupling the judging matrix: functionality and delivery are different
The Dreamforce Hackathon assessed technical indicators—user experience, agent capabilities and platform usage—separately from the presentation indicator of demo delivery.
- Define the demonstration scope in advance: The planning team must not give only a broad instruction to “present for five minutes.” So that judges can complete their score sheets, the rules need to require a three-stage scenario: the user’s first input screen, the core function in operation (processing), and the final output. An encyclopaedic presentation listing every app feature leads to overruns and boredom. Teams need guidance to show one compelling use case from start to finish.
2. Critical live-demo risks and failover rules
A live demo that runs the app directly on stage has a very high probability of failure. If the screen freezes or an error window appears during the demonstration, allowing a flustered host to delay the stage is the worst kind of on-site incident.
- Require a video backup: The day before the competition, the planning team needs to require every participating team to submit an MP4 recording of its demonstration running correctly, and have it ready on the console.
- Establish a retry cut-off: Clear rules are needed on how far a team may prolong a live broadcast by saying, “Just a moment, let me try again,” when its app stops. For example, require both participating teams and judges to sign off in advance on a failover rule: “One retry is allowed after a failure, with a maximum wait of 30 seconds. After that, the presentation immediately switches to playback of the backup recording.” This prevents disputes over fairness.
3. Control the handover between the stage console and presenter
The time a presenter spends connecting a personal laptop to the stage screen at the podium is a major cause of event delays.
- Make a tech run mandatory: Before the main event, connect every team’s laptop in the actual cue-sheet order to check resolution and aspect ratio, HDMI and USB-C adapter compatibility, audio output, and stage Wi-Fi access. If a device displays an incorrect aspect ratio or cannot mirror its screen because of security software, the organiser needs to make the decision to transfer that presentation environment to a spare shared PC prepared for the event.
4. A four-stage protocol for a hackathon demo stage
When planning a competition that includes IT technology and app demonstrations, the practical team needs the following four stages.
| Stage | Working standard |
|---|---|
| [Stage 1: Publish the demo-judging matrix in advance] | Transparently publish a judging framework that separates technical implementation from demonstration clarity and the UX scenario, beginning at the participant-recruitment stage. |
| [Stage 2: Collect failover videos] | Prepare for live-demo failure by mapping participants’ screen recordings, with subtitles or narration, to the VJ console in the correct order. |
| [Stage 3: Apply a device cut-off rule] | Limit delays in connecting a personal device on stage to a maximum of one minute. Agree a rule that, if the problem cannot be resolved, the presentation switches to launching on a shared PC or playing the backup video. |
| [Stage 4: Record and share incidents] | Have the host record technical actions taken on stage, such as a demonstration stopping, a retry or replacement video playback, with timestamps and share them with the judges in real time. This helps judges distinguish technical-infrastructure problems from a participating team’s mistakes. |
Planner’s note: a perfect demo comes from removing the freedom to fail
Hackathon participants often do not realise how helpless the code they have worked on all night can become when faced with the venue’s wireless internet or a projector’s resolution.
The planner’s role goes beyond believing in their enthusiasm and applauding it. It is to eliminate in advance every physical variable that could cause their demonstration to fail miserably on stage. The planner’s exacting work—checking cables, insisting on backup videos and imposing cool-headed time cut-offs—is the real consideration that delivers the participants’ hard-won results to the audience and judges as completely as possible.
Conditions for applying this elsewhere
The four criteria are published execution evidence. Live-demo, recording and retry procedures are proposals for another competition.
Information not established
Criterion weightings, team scores, feature tests, equal demo-time or retry rules and the winning app’s deployment impact were not established. Published criteria and awards do not verify fairness or product performance.
Reference material
- Salesforce official closing live blog · 16 October 2025: winning and runner-up teams, and the criteria used in judging
- Devpost operating partner’s retrospective: competition for Dreamforce 2025 attendees, winning app and grand prize
- Salesforce Developers official event guide · September 2025: San Francisco conference dates, 14–16 October, and the hackathon presentation programme
Have a correction or additional evidence? Request a correction
