More founders now start alone. Carta’s Solo Founders Report 2025, based on data from tens of thousands of US companies, shows solo founders rising from 23.7 per cent of new startups in 2019 to 36.3 per cent in the first half of 2025. The same report shows the price: solo-led companies made up 30 per cent of the startups founded in 2024 but raised only 14.7 per cent of the cash in priced equity rounds that year.
For accelerators, university entrepreneurship centres and startup hubs, that gap is a programme question. Many founders who apply to you alone would rather not stay alone. The usual answer is a co-founder mixer on a weekday evening: drinks, name badges, a few pitches. It fills the room. It rarely forms a team.
A co-founder matching event works when you match on role and commitment before anyone arrives, and when you treat the evening as the first of several meetings. Nobody chooses a co-founder in one round of conversation. This guide shows which formats exist, why open mixers underperform, what to match on, how to fix the usual shortage of technical founders and how to measure the result in teams rather than contacts.
What is a co-founder matching event?
A co-founder matching event brings together people who want to start a company and are looking for a partner to do it with. Some arrive with an idea and need a team. Others bring skills such as engineering, design or sales and want to join an idea. The host is usually an accelerator, a university, a startup hub or a community group, and its job is to make the right people meet. Three formats dominate:
| Format | How it works | Example | Strength | Weakness |
|---|---|---|---|---|
| Open mixer | Free networking, often with visual cues for roles and short pitches | Tufts Cross-University Co-Founder Matching Mixer, April 2026: one role sticker (Asker, Hacker or Hustler), one skill sticker, rotating stations, 60-second pitches | Low effort, many first contacts | Who meets whom depends on who approaches whom |
| Structured rounds | Pre-arranged one-to-one conversations in fixed rounds, based on profiles | Munich Co-Founder Matching, April 2025: two 45-minute matching rounds. imec.istart (online): profiles with requests and offers, invitations a week ahead, 12-minute video calls | Every participant meets relevant people | Needs profile data and planning |
| Project sprint | Teams form around pitched ideas and build something together | Techstars Startup Weekend: 54 hours, pitches on Friday, teams of strangers present on Sunday (Tufts Derby Center, May 2026) | Shows how people actually work together | High commitment, fewer participants |
The formats work best as stages rather than alternatives. A mixer finds candidates, structured rounds test fit in conversation, and a sprint tests fit in real work. Universität Hamburg’s Startup Port runs its Meet&Match as an accompanied process over several weeks, with limited places, three times a year. That is closer to a small programme than to an evening event.
Why open mixers rarely produce teams
The open mixer is still the most common format, and it has a structural weakness. People talk to people they already know, or to people like themselves.
Research on founding teams has shown this for decades. Ruef, Aldrich and Carter analysed 816 founding teams from the US Panel Study of Entrepreneurial Dynamics (American Sociological Review, 2003). The strongest pattern in team composition was homophily: founders teamed up with people of the same gender, ethnicity and occupation. Functionally diverse teams were no more common than chance would predict. Left alone, engineers found companies with engineers, and business graduates with business graduates.
A field experiment adds the event angle. Sharique Hasan and Rembrand Koning followed 112 aspiring entrepreneurs at a three-week startup bootcamp in New Delhi and assigned teams at random (Strategic Management Journal, 2019). Participants who knew nobody beforehand had a 25 per cent chance of interacting with a given peer. For participants who already knew others, it was 16.6 per cent. “Once you have pre-existing ties, you just explore less,” Hasan summarised (Duke Fuqua Insights, 2019).
At a university mixer, prior ties are the default. Half the room arrives with classmates. They stand together, they leave together, and the computer science student who came alone ends up talking to the bar staff. A free room reproduces existing circles.
Now picture the programme manager of a university founders’ centre the morning after. Eighty people came, the photos look good, and the dean asks how many teams formed. She has no answer. Nobody will give her one for months.
What to match on: role, commitment and idea status
If the room will not mix itself, the organiser has to decide in advance who should meet whom. The most detailed public data on what founders look for comes from Y Combinator’s (YC) online co-founder matching platform. Three months after launch, it had 16,000 profiles, 130,000 invitations and 33,000 matches:
Co-founder preferences
What founders look for in a co-founder
Each of these preferences translates into a profile question and a matching rule. Five attributes carry most of the weight:
- Role: Technical, business, design or domain expert. The default rule pairs complementary roles, technical with non-technical, because that is what most matches look like.
- Time commitment: Full time now, full time within six months, or part time. A student with exams until July and an employee ready to resign next month are not a match, however well they get on.
- Idea status: Has an idea, open to ideas, or runs an existing startup and wants a late co-founder. Pair idea holders with open people; two idea holders tend to spend the meeting defending their own idea.
- Problem interest: Health, climate, fintech or another field. In the YC data, 72 per cent of matches shared at least one interest.
- Start date: When the person wants to start. Of the YC founders not yet working on a startup, 47 per cent planned to start within three months.
These are explicit rules, not a black box. You can tell every participant why they are meeting whom, and you can change a rule when the mix of people changes. Which of these questions belong in the registration form and which come after confirmation is covered in our guide to registration questions for event matchmaking.
Fix the supply of technical founders before you open registration
Matching rules only work if both sides of the match turn up. In co-founder events, they rarely do in equal numbers. On the YC platform, 42 per cent of users identified as technical, while 62 per cent of founders said they wanted a co-founder who does engineering. Demand for engineers exceeds supply even on a platform built in the software world. At a mixer hosted by a business school, the gap can be much wider.
This is a recruitment problem, and it starts weeks before the event:
- Co-host with computer science, engineering, science and design faculties, not only with the business school. The Tufts mixer draws on five institutions, including an art and design school.
- Set a target ratio per role and cap registrations for the overrepresented role. Put the rest on a waiting list rather than overfilling the room.
- Give technical participants a reason to come: share the list of ideas and profiles in advance, as imec.istart does one week before its online event.
- Make “open to ideas” a first-class profile, so engineers without their own idea do not feel like extras.
Ten engineers and sixty idea holders do not make an event. They make a queue.
Plan the evening as the first of several meetings
Even a perfectly balanced room cannot answer the real question in one evening. Choosing a co-founder is closer to a hiring process than to a networking contact, and new teams tend to rush it. Noam Wasserman’s research on nearly 10,000 founders of US technology and life sciences startups found that 73 per cent made their equity split decision in the first month (Wasserman, The Founder’s Dilemmas, 2012). Teams that form at your event may make their most consequential decision before they know how they work together.
Your programme can slow that down by building the next steps into the format:
Process
From profile to trial project
- Before Profile and commitment check Collect role, commitment, idea status and start date, build the meeting plan and share profiles a week ahead
- Evening Two to three structured rounds Short one-to-one conversations of 12 to 20 minutes with breaks, then open networking
- Week 1 Second meeting Pairs who both want to continue get a reserved slot within a few days
- Weeks 2 to 6 Trial project A small joint task, customer interviews or a weekend sprint before any decision on equity
Timing matters for the second step. In the YC data, half of the matched pairs who met did so within 6.2 days. Interest fades quickly, so book the follow-up slot while it is still there. For the trial project, a short build format works well; our guide on how to host a hackathon covers the planning.
One thing the organiser should not do is advise on equity splits, vesting or founder agreements. Invite a lawyer or an experienced founder to give a short talk on what such agreements usually cover, and point teams to professional advice.
How co-founder matching differs from mentoring and community events
This follow-up logic resembles mentoring, but the matching problem is different. In mentor office hours, the relationship is asymmetric: founders ask, mentors give, and demand piles up on a few famous names, so the main task is capping their slots. Co-founder matching is symmetric. Every participant is both seeker and candidate, both sides must say yes, and the scarce resource is a profile type, usually the engineer, not a person.
It also has a place in the wider calendar. In our guide to ecosystem builder events, co-founder matching fits the quarterly layer: too demanding for a weekly meetup, too important to leave to the annual startup week.
Solution: Converve schedules pre-booked one-to-one meetings with a rule-based meeting matrix. You define the profile attributes, such as role, commitment and idea status, and the rules for which groups should meet. Participants then request and accept meetings within those rules, and follow-up meetings for the same pairs can be scheduled after the event. You can see how this works for startup and investor events.
Measure teams, not contacts
The programme manager from the morning after needs a different set of numbers. Attendance and contacts exchanged say little about whether the event did its job. Four measures tell her more:
- Share of participants with at least two meetings that matched their rules.
- Number of pairs who both wanted a second meeting, and how many held it within a week.
- Number of teams still working together after 90 days.
- Number of those teams that apply to your accelerator, incubator or a founder grant.
The last two numbers take three months to collect. They are also the only ones the dean will remember.
Frequently Asked Questions
What is a co-founder matching event?
A co-founder matching event is a format in which an accelerator, a university, a startup hub or a community group brings together people who want to start a company and need a partner. Typical formats are open mixers, structured one-to-one rounds based on profiles, and project sprints such as a 54 hour Techstars Startup Weekend.
How do you find a technical co-founder at an event?
Look for events that recruit beyond the business school and match by role. At the event, state clearly what you offer, for example customer access or domain knowledge, and be open about your commitment. Engineers are in short supply: on Y Combinator’s platform, 42 per cent of users were technical, while 80 per cent of non-engineering founders wanted an engineer (YC, 2021).
How long should conversations at a co-founder matching event be?
For structured one-to-one rounds, 12 to 20 minutes per conversation works well, with a short break between conversations. imec.istart, for example, runs 12-minute video calls with three-minute breaks. Longer conversations belong in the second meeting, once both sides know they want to continue.
Do online co-founder matching events work?
They can, especially for participants from several cities. Half of the users on Y Combinator’s platform said location did not matter to them, yet 55 per cent of matches were in the same country (YC, 2021). Online rounds work well for a first contact; the second meeting and any trial project benefit from meeting in person.
Should new co-founders sign an agreement straight away?
Co-founders should agree on roles, commitment and equity in writing, but only after they have worked together for a while. Wasserman’s research found that 73 per cent of founders split equity in the first month (Wasserman, The Founder’s Dilemmas, 2012). Organisers should not give legal advice; a short talk by a lawyer and a referral to professional advice is the right level.
Conclusion: the evening is the start, not the result
A co-founder matching event succeeds when the organiser does the matching work before the evening and plans the meetings after it. Match on role, commitment and idea status, recruit enough technical founders to balance the room, and give every promising pair a second meeting and a trial project. A free mixer produces contacts. A planned sequence produces teams the programme manager can report.
If you are planning a co-founder matching event, a founder programme or investor meetings at your accelerator or university, get in touch with Converve. We will show you how a rule-based meeting matrix and follow-up meetings fit into one programme.
