’Twas the week before Valentine’s, when all through Harvard, singles were stirring, with hope in their hearts.
Their survey answers were entered into Datamatch with care, in hopes that finding love soon would be there.
The lonesome were nestled, all snug in their beds, while a computerized cupid “Algorithm™” toiled, several steps ahead.
Harvard students might go all four years without finding a special Harvard someone. Yet, it takes less than seven days for Harvard’s very own data-driven solution to provide an opportunity to not spend Valentine’s Day alone. The rumors are true—Datamatch has returned for its 29th season of love. An online matchmaking service, Datamatch uses an “Algorithm™” to suggest matches for users based on their desire to find “love” or “friends.” Available a week before Valentine’s Day, students at Harvard and 50 other colleges can create a profile and fill out a survey of lighthearted questions. The “Algorithm™” runs once the survey closes at 11:59 p.m. EST on February 13th, and students wake up on Valentine’s Day to find the resulting matches in their inbox.
The original Datamatch was launched on Harvard’s campus as a paper survey in 1994, one year before the launch of the pioneering online dating website, Match.com. In the 30 years since its analog beginnings, Datamatch has seen significant development in both its popularity and algorithm. “The algorithm gets quite complicated, hence, they use machine learning techniques to create matches,” stated one of this year’s Supreme Cupids, Nadine Han ’25. As a Supreme Cupid, Han shares a co-president-like role in the Datamatch organization with Lily Liu ’25.
The “Algorithm™” matches students at the same college based on their survey answers, profile information, and other Datamatch site interaction. Of the matches suggested for each user, they can select who they “match” with, i.e., who they are interested in meeting up with. A number of sponsored meals are available for mutual matches to take advantage of at nearby restaurants, an enticement for users to go through with seeing where it leads.
Most dating apps let users control their interactions with others, based on their preferences. According to Supreme Cupid Han, Datamatch differs from this model since users can personalize their profile and survey answers, but it is the “Algorithm™” that selects matches. “We pick people for you to interact with,” she explained. “This idea of having a customized algorithm that decides what might be best for you…is also really exciting and different.”
Han thinks Datamatch has become popularized at Harvard for “this concept of casual matchmaking slash low stakes opportunities to meet other people.” In 2023, 60% of Harvard students filled out the survey, according to Han. “Harvard is the highest percentage, by a good margin” stated Howard Huang ’26, Datamatch’s Stats Lead. “I filled out the survey because everyone else was doing it,” explained Tali Wong ’25.
Amid the developing world of artificial intelligence and data science, it is no surprise that data-driven problem-solving now extends to the dating realm. “Technology has gotten to a point where we can leverage them to find partners,” said Mayesha Soshi ’24, a Teaching Fellow for the Harvard course Critical Thinking in Data Science. Students in this course examine algorithmic biases and ethical considerations in data science models and data collection.
People tend to romanticize algorithm-based dating platforms, said Queenie Luo, PhD student and Teaching Fellow for Critical Thinking in Data Science. “They think algorithms and math is more objective,” she stated. “They think data science can help them to reach a broader range of people.”
The downside of all of this? The scope of the actual algorithm. “The Datamatch survey is very rudimentary. It’s not complex at all,” said Soshi, having taken the survey in the past. She voiced that dating algorithms are dependent on only the data available. “There’s so many things to consider when it comes to dating,” she said. “You need to be able to use numbers to quantify things, and you’ll lose a lot of information in the process.”
“We live in a digital age where everything is very heavily focused around social media and the internet,” said Han. “On one end, it might seem a bit impersonal, because it’s not a human who’s making these decisions for you. But I think on the other end, you’re able to reach a lot more people.”
Soshi also recognizes the lighthearted nature of Datamatch. “It’s a fun little activity for college students to engage in,” she said. “You can find friends, you can find potential romantic matches, and you might go on that free meal day and it might be a fun story.”
Despite the draw of Datamatch and its success stories, it inevitably does not work for everyone. “No one matched me back,” said Wong about her experience with Datamatch last year. According to Huang, 33,338 of the 179,842 algorithm-suggested matches last year resulted in mutual pairings, or roughly 18.5%. Datamatch cannot currently tell whether a mutual match goes through with a free food outing.
When considering why students may not follow through with a match or take advantage of the opportunity for free food, Wong thinks that finding a true match may not have been an intention in the first place. “People fill it out for the curiosity of who they’re going to match with and less for actually looking for something,” she said. “At Harvard, people are busy. Free food is exciting, but going on a date with someone you don’t know could be kind of awkward.”
Ultimately, the Supreme Cupids envision Datamatch as a fun activity for Harvard students. “Datamatch is much more lighthearted and not super serious,” said Supreme Cupid Liu. “I would say there is some value in having some sort of technical algorithm contribute to how people get matched together, although only up to a certain extent.”
“We’re here for the student body,” said Han. The team works to present Datamatch in an appealing light. “The less intimidating this type of application is, I think the more likely people are to engage in it,” she said. “The tone of the website is very light-hearted and I won’t say goofy, but right now, it’s a video game retro theme,” said Liu. In this fun spirit, the survey contains twelve multiple choice questions with five comical answer choices. For example, in response to “What’s your Roman Empire?” a user could respond that it is “My p-set buddies to lovers arc fantasy.” To see more questions, fill out the survey yourself!
Beyond providing an opportunity for Harvard students to connect, one of Datamatch’s goals as an organization is to teach compers about coding languages and technical tools. A fall semester boot camp prepares Datamatch Team hopefuls to contribute to the site in various capacities: Algorithm, Statistics, Design, Business, and Web. “We’re just a bunch of people who take a semester to learn something new. And then we kind of apply it to this huge project that has thousands of people on it,” said Han.
The hard-working team and popularity behind Datamatch make it an enduring annual activity. “It has kind of become a Harvard tradition at this point,” said Liu.
Clara Lake ’27 (claralake@college.harvard.edu) wonders how her first experience with Datamatch will go.