Few words capture the spirit of building something new quite like “hackathon.” It signals a particular kind of belief—that given a little time and freedom, people will create something remarkable. But in my experience, the reality never lived up to the promise.
At YouTube, we held a big annual hackathon, and the unspoken rule was simple: You were expected to “ship” something, or build and launch something functional.
This mentality pushed people toward familiar ideas and familiar teammates. In the end, we had a collection of prototypes, but the experience was neither energizing nor inspiring for the team.
When I founded Coda, my instinct was to skip hackathons. But one of our engineers was passionate about their value. We decided to try again with one rule that changed everything: nothing ships.
It turned out to be more powerful than expected. Over the years, we ran over 20 hackathons at Coda, and a surprising number of our most important ideas began in those sessions. Removing the pressure to ship gave people the freedom to be curious. It is an approach I continue to carry with me at Superhuman, where the same principles help our teams explore more boldly.
The power of not shipping
Innovation requires vulnerability. It asks people to propose half-formed ideas, to explore areas they do not fully understand, and to work with people they do not usually team up with. When a hackathon is tied to shipping, people instinctively narrow their choices. They gravitate toward projects that feel achievable and seek out familiar teammates who can help them get there quickly.
Once shipping is removed from the equation, people explore ideas they might normally set aside. They bring in colleagues from outside their usual circles. And people who are not engineers begin to participate with real confidence. The output of the hackathon becomes the idea itself, not a finished product.
Ideas also take time to grow into their best form. One of my favorite examples, Coda’s Buttons feature, surfaced in four separate hackathons before becoming part of the product. With each hackathon, the concept shifted and expanded, and no one team was trying to finish it. What emerged became one of Coda’s most powerful features, which was the ability to take action directly from a doc with a single click, automate tasks, and trigger workflows. Many breakthroughs arrive this way, in layers rather than leaps.
Fast, frequent, light hackathons
A common debate about hackathons is frequency. Many teams start with the premise that “we take this seriously, so we dedicate a whole week to it.”
But this creates a perverse incentive: the longer the hackathon, the more pressure participants feel to “show results” that justify the “displaced work.” Teams with longer hackathons also tend to run them less frequently—often just once a year—which makes it harder for the ritual’s mechanics to take shape and creates even more pressure to ship.
I’ve found the most success with short hackathons of a single day or two. Especially in the AI era, a small team can produce amazing results in a day that capture 80% of an idea. If the goal is the idea, then 80% is usually a good stopping point.
A 24-hour ritual is also easier to maintain frequently, such as every quarter. I met a company that even runs weekly hackathons, but just for three hours on Wednesday evenings. Fast and frequent turns out to be a winning combination.
Team size matters too. Two to seven people tend to create the right mix of diversity and momentum. Solo projects are discouraged—if you can’t convince at least one other person to join you, that’s a useful signal.
And larger teams are encouraged to split up. A ten-person team will spend most of its time just getting aligned. With only 24 hours, you must work fast.
Hackathons create permission
Across Google, Coda, and Superhuman, I often hear leaders say they wish their teams had more ideas. But in my experience, the issue is more about permission.
People have ideas they care about, but lack the space and psychological safety to explore them. Hackathons create that space. They give people a moment to step away from their day-to-day work and chase something simply because it interests them.
At our last Superhuman hackathon, over 500 participants generated 278 ideas. People felt free to say, “Here is something I have been thinking about.” When teams began working, those ideas naturally evolved, shifted direction, and sharpened as new insights emerged. I see this not as indecision, but learning in motion. The best hackathons normalize that evolution.
Hackathons also deepen connections across the company. A marketer partnering with an engineer. A designer working with someone in customer support. These pairings rarely happen in day-to-day work, yet they often lead to the most memorable outcomes.
And yes, the swag somehow matters. A good T-shirt lifts the whole event more than I can explain.
5 tips for leaders to host a great hackathon
If you’re thinking about bringing hackathons into your organization, here are practices that have consistently worked across the teams I’ve led:
- Set the rule that nothing ships: Removing the pressure to produce opens the door to risk-taking and honest exploration.
- Make it inclusive: When a wide range of people are involved, the ideas become richer.
- Encourage small teams: Two to seven people create the right mix of diversity and momentum.
- Keep it short and run it often: A 24-hour format each quarter is an easier ritual to sustain than a week-long event run annually.
- Celebrate exploration, not competition: The goal is to surface possibilities, deepen connections, and strengthen the culture.
In the end, a great hackathon is about giving people room to wander, to pursue an idea that has been sitting at the back of their minds, and to reconnect with the joy of building something new. With the right structure and the right spirit, hackathons can become one of the most effective rituals an organization has for sparking innovation.
Shishir Mehrotra is the CEO of Superhuman.