How Hackathon Judging Works: Criteria, Panels and Scoring

hackathon judging criteria

Hackathon judging criteria are the scoring categories judges use to compare projects built in a short event, most often the quality of the idea, the technical build, design and potential impact. Each organiser sets the final list.

Judging sits at the end of every hackathon, including the remote events that grew during the covid-19 hackathon wave, so it helps to know what is being measured before a team writes a line of code.

The criteria judges score most often

There is no single universal rubric. Devpost, the platform that hosts a large share of online hackathons, says that hackathon judges score submissions against criteria set by each hackathon’s organizers. The same Devpost guide adds that the list typically includes technological implementation, ease of use, demonstration, potential impact, quality of the idea, and design.

Devpost’s organiser help centre narrows that pool to a short core. It says the most common criteria include the quality of the idea, the implementation of the idea and potential impact, and it describes the first as a question of how creative and unique the project is. The implementation criterion asks whether the end result demonstrates quality software development and design.

  • Quality of the idea: is the concept original, and does it address the theme the organisers set?
  • Technological implementation: does the prototype work, and does it use the required tools well?
  • Design and ease of use: can a stranger understand the project in a sentence or two and use it without help?
  • Potential impact: does the project solve a genuine problem for a real group of people?
  • Demonstration: does the demo or video show the project working and explain it clearly?
hackathon judging criteria scoring sheet
A clipboard with a colourful scoring chart lying on a wooden table.

How weighting changes the scores

Criteria only matter once they carry weight. Devpost’s help centre says that judging criteria is usually weighted equally and this is how the Devpost judging platform is set up. Organisers who want a different balance, for example more weight on impact for a civic or social-good theme, have to judge offline and publish the percentages themselves.

The same page tells organisers that they should define the criteria and their weight before launching the hackathon. That advice protects participants as much as organisers: a team that knows impact counts for more than polish will spend its limited hours differently. The Devpost help article was last updated on November 1, 2024, so readers should check the current help centre for the platform’s present behaviour.

Criteria also reflect the culture of the event. Major League Hacking, the student hackathon league, tells organisers to remind judges about the criteria and not to focus on the business side of the event, a point it makes in its guidance on recruiting and briefing judges. In practice that means a polished product pitch should not beat a clever technical hack on charm alone.

Who sits on a judging panel

Most events use one panel of subject-matter professionals who all score the same criteria. Devpost notes that category prizes sometimes involve separate groups of experts, and it advises that each panel is made up of at least three diverse and co-ed individuals. To limit confusion it recommends limiting yourself to 2 – 4 panels.

Public voting is a separate path. Devpost describes community-choice prizes as a way to widen publicity, while warning that fraudulent votes are a risk and recommending small prize values and holding back live results until votes have been reviewed.

The judges themselves tend to come from the organisers’ networks. MLH’s recruiting guidance lists school professors, deans and alumni, local community leaders and representatives of local technology companies, and it advises that the judges chosen have strong technical aptitude.

Science-fair judging, step by step

For in-person events, MLH is direct about the format it prefers. Its judging plan states, Our strong recommendation is implementing a science-fair type of judging plan. Teams set up at numbered tables, judges walk the room, and every team gets to present to people who can ask questions.

MLH argues against sending teams to separate judging rooms. In its words, those format types frequently have a hard time being fair because teams are only judged against the projects that happened to have the same room. A team seen in only one room has to be re-pitched at the final stage, which depends on judges’ memory and notes.

The same plan gives rules of thumb for the schedule. MLH recommends 3 rounds per project and 4 mins per project for each judge, and it says judging usually takes around 2 – 3 hours. Judges are briefed that they have only 3 minutes per team, so that the timeline stays fair across the room. Those figures come from the MLH judging plan and are guidance for organisers, not a rule that every event follows.

Stack ranking and choosing winners

Raw scores have a known weakness: some judges are generous and others are strict. MLH’s answer is stack ranking, which it prefers because it normalizes scores across judges. Each judge reports back their top 3 projects from the set assigned to them, and the organisers award 3 points, 2 points and 1 point for first, second and third place. The projects that appear most often near the top rise to the shortlist.

The shortlist is then checked by people, not just a spreadsheet. The plan tells organisers to look at the leading few projects and make a final decision, so that the result is one they can stand behind. Category prizes work differently: for narrow tracks such as hardware, MLH suggests one or two subject-matter experts who view only that track’s projects.

Criteria and ranking also depend on the submission itself. A demo video, a link to the code and a short description give judges something to refer back to during deliberation, and the MLH rules guidance tells organisers to decide in advance what has to be submitted and in what format.

Checks before winners are announced

Fair scoring assumes the work is the team’s own. MLH advises that organisers should still check all winning projects regardless of event type for cheating. A suspicious sign is a prompt for a conversation, not an automatic disqualification: the guidance asks organisers to talk to the team and ask who worked on which part.

It also reminds organisers that beginners make innocent mistakes. A repository with a single commit near the end can simply mean the team is new to version control, and MLH says that one commit at the end of the event is NOT a reason to DQ. Rules about starting work before the event, reusing earlier projects and crediting AI tools are meant to be stated clearly beforehand so that any decision can point to a specific rule.

Judging when the event is online

Remote events keep the same criteria but change the logistics. Judges review recorded demos or short live calls instead of walking between tables, and time-zone differences affect when scores can be collected. The guide to running an online hackathon covers those platform and scheduling choices in detail.

For participants, the lesson is to make the demo self-explanatory, because a judge may see it without being able to ask questions. Teams that divide the work well tend to have a cleaner demo, which is where understanding hackathon team roles pays off before judging starts.

This site is an independent editorial guide and does not run, score or register entries for any event. Rules, weights and prize details differ for every hackathon, so the event’s own page is the final word on how its projects will be judged.

Frequently asked questions

Are hackathon judging criteria the same everywhere?

No. Organisers set their own criteria. They tend to draw from a shared pool covering the idea, technical implementation, design, ease of use, demonstration and potential impact, but the final list and its weights are published on each event’s own page.

Are all criteria weighted equally?

Often they are. Devpost’s help centre says criteria are usually weighted equally on its platform, and that unequal weights require judging offline with the percentages announced to participants.

Why do some hackathons use stack ranking instead of scores out of ten?

Stack ranking asks each judge for a short ordered list instead of a numeric grade. MLH prefers it because it normalizes scores across judges, so a generous judge and a strict judge influence the result equally.

Can a project be disqualified after winning?

Yes, if a check finds that the rules were broken. MLH advises organisers to check winners before announcing them and to talk to the team first, since many warning signs have innocent explanations.