← Blog

Guides

What Judging Criteria Should a Hackathon Use?

Abhinavtej ReddyNovember 14, 20265 min read

A good hackathon rubric has three to five criteria. Each one needs a short written meaning and a fixed weight. Technical implementation, innovation, impact, design and presentation appear most often across published rubrics. The right mix depends on what your event is for [1][5][6].

The criteria most rubrics use

Organisers set their own criteria. Still, submissions are most often scored on technical implementation, ease of use, demo, potential impact, idea quality and design [1]. The weights vary a lot between published examples:

CriterionWhere it appearsWeight in the example
Technical implementationCommon in lists [1], HackHQ sample [5]25% in HackHQ's sample, raised to 35% for technical events [5]
Innovation or originalityNexHacks [4], HackHQ [5]25% in NexHacks, 30% in HackHQ
ImpactBellingcat [2], NexHacks [4]25% in NexHacks, 25% in HackHQ
Design and UXNexHacks [4], HackathonHost [6]15% in NexHacks, 20% in HackathonHost
Presentation or demoNexHacks [4], HackHQ [5]10% in NexHacks, 20% in HackHQ

The sources do not agree on one set. HackathonHost suggests impact, technical execution, design and UX, theme fit and presentation at 30, 25, 20, 15 and 10 percent [6]. HackHQ's sample has no separate design or theme criteria [5]. Treat any published set as a starting point. It is not a standard.

Bellingcat's rubric uses four areas: imagination, technical innovation, a working solution and potential positive impact [2]. It also asks whether a similar tool already exists. It asks how complete the tool is, compared with how much it relies on mockup data [2]. That last question matters in any hackathon where demos are easy to fake.

Weight by event goal

The weights should follow the reason you run the event. The Hacktribe blueprint raises impact and feasibility to 30 to 35 percent each for corporate sprints. It raises technical execution to 35 percent for deep-tech or AI events [9]. A first college hackathon with a technical brief will usually lean toward technical execution and a working demo. A sponsor-led challenge will usually lean toward impact on the sponsor's users.

Write down the reason for each weight. A weight with no stated purpose gets argued about on the day.

Keep it to four or five criteria

Marachallenge advises three to five criteria. Past five, each one weighs little and final scores look alike [7]. HackathonHost advises four to six. It warns that more than six makes judges lose focus, and fewer than four makes scores too blunt [6]. The overlap is four or five. That keeps the sheet short enough to use on the day, without any single criterion weighing too little.

Make every criterion specific

A bare word like "impact" says nothing. Marachallenge recommends naming who or what each criterion affects [7]. Compare these two versions:

  • Vague: Impact, out of 5.
  • Specific: Impact on first-year students who miss deadlines, judged on whether the team shows a clear path to real use by that group, out of 5.

The second version tells a judge what to look for. It also tells teams what to build toward. The DEV Community argues that terms such as "innovation" or "impact" are ambiguous. Judges apply unwritten personal rubrics, so teams may build for what they guess will score [8]. Writing the meaning down fixes this.

Decide what you will not score. Bellingcat says code quality and documentation play no role in its judging. It advises organisers not to spend time on them [2]. Your rubric can make the same choice, but say so in the brief.

Use one scale and one meaning per score

CGU's ethical AI rubric scores each criterion from 1 to 5 [3]. NexHacks does the same [4]. HackHQ suggests 1 to 10 for most events, and 1 to 5 for small, simple ones [5].

Marachallenge's method is to score each criterion from 1 to 5, divide by 5, then multiply by its weight. That gives a total out of 100 [7]. For example, a hypothetical team scores 4 out of 5 on technical execution, with a 30 percent weight. It earns 0.8 times 30, which is 24 points toward the total.

Give every judge the same rubric and a one-line meaning for each score. HackathonHost notes that consistency between judges matters more than any one judge's expertise [6]. A panel that agrees on what a 3 looks like will produce fairer results than experts who each use their own scale.

Avoid criteria the brief never asked for

Marachallenge advises against judging criteria the brief never asked for. Financial feasibility is one example, when the brief says nothing about costs [7]. If your problem statements ask for a working prototype, score the prototype. Do not add a business plan score that teams were never told to build.

Where Unifesto Sprint fits

If you want weights and scores in one place, Unifesto Sprint supports hackathon teams, problem statements, rounds with weighted rubrics, judge panels, a live leaderboard and certificates. It is part of the Pro plan. Software cannot decide what your event values, so the choices above still come first.

For the wider setup, read how to organise a college hackathon. It covers the format, the budget and publishing the rubric before teams start. For panel size, see how many judges you need for a hackathon. If judges are seeing pre-built work, how to stop teams submitting pre-built projects explains how to combine rules with judging in rounds.

Our view

Our view: for a first college hackathon, use five criteria: technical execution, impact on a named group, originality, a demo that works, and design. Give technical execution the largest weight. Publish the one-line meaning for every score, and do not score anything the brief did not ask for. Publishing the rubric early gives teams a clear target to build toward.

Limitations of this advice

Most sources here are vendor blogs and guides, and the weights they suggest conflict with each other. We found no India-specific judging guidance, such as rules for government or college hackathons. Check your college's own requirements. Several sources were read only through search snippets, not full pages, so the details above may be incomplete.

FAQ

How many judging criteria should I use?

Use three to five, and four or five if you can. Past five, each criterion carries so little weight that final scores look alike [7].

Should I score code quality?

Only if your brief says so. Bellingcat does not score code quality or documentation at all [2]. Marachallenge advises against judging anything the brief never asked for [7].

What scale should judges use?

A 1 to 5 scale per criterion is common [3][4]. Give every judge the same meaning for each score, so their marks are comparable [6].

Sources

  1. Understanding hackathon submission and judging criteria: What really counts: https://info.devpost.com/blog/understanding-hackathon-submission-and-judging-criteria
  2. Bellingcat Hackathons Grading Rubric August 17, 2022: https://www.bellingcat.com/app/uploads/2022/08/Bellingcat-Hackathon-Grading-Aug22.pdf
  3. JUDGING RUBRIC - Ethical AI Hackathon: https://research.cgu.edu/hackathon/home/judging-rubric/
  4. Judging Criteria: https://www.nexhacks.com/hacker-resources/judging-criteria
  5. How to Judge a Hackathon: Fair Criteria & Rubrics: https://hackhq.io/planning-guide/judging
  6. Hackathon judging criteria & scoring rubric, HackathonHost: https://hackathonhost.com/articles/hackathon-judging-criteria
  7. Hackathon judging criteria: 3 ready-to-use rubrics: https://www.marachallenge.com/en/guides/hackathon-judging-criteria
  8. When Judge Bias Matters More Than Eligibility: The Hidden Problem With Modern Hackathons - DEV Community: https://dev.to/sumit0rn/when-judge-bias-matters-more-than-eligibility-the-hidden-problem-with-modern-hackathons-1h73
  9. Fair & Balanced: The Hackathon Judging Rubric Blueprint: https://hacktribe.co/blog/fair-balanced-the-hackathon-judging-rubric-blueprint