How VC Funds Find Startups Before They Fundraise
Venture investors find companies before they raise by watching for the work that comes first: a founder leaving a job, a company being incorporated, a domain registered, code pushed to a new repository, a first hiring post. They then check each signal against their own thesis and reach out with a reason to talk, before a round opens and the company has a queue of investors.
The rest of this post covers where deal flow actually comes from, which early signals carry information and which mostly carry noise, and a six-step workflow for going from signal to first conversation.
Where venture deal flow actually comes from
Most venture deals do not arrive cold. In a survey of 885 institutional venture capitalists at 681 firms, Paul Gompers, William Gornall, Steven Kaplan and Ilya Strebulaev asked where their deals originate (NBER working paper 22587, Section 3.1):
| Source of deal | Share of deals |
|---|---|
| Professional network | Over 30% |
| Proactively self-generated | Almost 30% |
| Referred by other investors | About 20% |
| Referred by a portfolio company | 8% |
| Inbound from company management | About 10% |
Two things stand out. Only about one deal in ten came in cold from the founders. And almost as many deals were generated by the investors' own outbound work as came through their networks. Finding a company early is not a niche tactic; it is a large share of how venture firms already source.
The same survey found firms considered roughly 100 opportunities for every deal they closed. Earlier discovery does not shrink that funnel. It changes your position in it: you meet the founder as someone who noticed their work, not as one of many term sheets.
What "before the fundraise" actually means
The window runs from the first visible sign of a new company to the moment the founder starts taking investor meetings. That can be weeks or many months. It is when a founder is deciding who to talk to, has time for a real conversation, and has not yet anchored on a valuation or a lead.
The goal in that window is a relationship and context, not a term sheet. An investor who has followed a team's progress for three months can move with conviction when the round opens. One who meets them at the round cannot.
The early signals investors watch, and what each can tell you
No single signal proves a company is worth backing. Each one answers a narrow question, and each needs checking before anyone acts on it.
| Signal | What it can suggest | What it cannot tell you | How to check it |
|---|---|---|---|
| Founder leaves a role | Someone may be starting something | Whether they are, or what it is | Their public profile, prior work, who left with them |
| Company formation or incorporation | A legal entity now exists | Ambition, sector or venture scale | Registry record, founders named, address |
| Domain registration | A name and product are being prepared | Whether anything ships | Live site, registration date, linked accounts |
| Code activity in a new repository | Something is being built | Product quality or demand | Commit history, contributors, what the code does |
| First hiring posts | The team is growing and has some funding or conviction | Revenue or traction | Roles posted, seniority, location |
| Founder public activity | Thinking, audience and direction | Execution | Posts, talks and writing over time, not one post |
The signals are strongest in combination. A founder who left a payments company, incorporated a company, and opened a repository with ledger code in the same month tells you far more than any one of those events alone.
Why raw signals are mostly noise
Company formation alone is far too broad to act on. The U.S. Census Bureau counted 534,947 business applications in September 2026 alone, seasonally adjusted (Business Formation Statistics, released 14 October 2026). Almost none of those will ever raise venture capital.
That is why every useful sourcing process has a filter, and the filter is the fund's own thesis: the sectors, founder backgrounds, stages and markets it actually backs. Two funds can see the same new company and reach opposite conclusions, and both can be right. A signal feed without that filter just moves the noise from one inbox to another.
A six-step workflow from signal to first conversation
- Write the thesis as signals. Turn "we back infrastructure founders from fintech" into things you can observe: prior employers, technical roles, the kind of repository or domain a team in that space would create.
- Collect signals from sources you trust. Decide which sources count and record where each signal came from, so anyone on the team can check it later.
- Qualify before anyone reviews. Set a minimum bar of evidence (for example, two independent signals plus a relevant founder background) before a company reaches an analyst. This protects the team's time.
- Save it to a radar and watch for change. Most early companies are not ready for a conversation the day you find them. Track them and let new signals tell you when the picture has moved.
- Reach out with a reason. Reference the specific work you noticed and why it fits what you back. A message that could be sent to any founder will be treated like one.
- Log the outcome. Record whether you met, passed or were too early, and why. That history makes the next decision faster and keeps the firm from researching the same company twice.
Common mistakes
- Treating a signal as an opportunity. A signal starts an investigation. It does not show investment fit.
- Scoring everything the same way. A generic score hides why a company was surfaced. Keep the reasons visible.
- Reaching out with nothing to say. Early outreach works when it shows you understand what the founder is building.
- Ignoring time. Signals go stale. A domain registered two years ago with no site says something different from one registered last week.
Doing this with software
Much of steps 2 to 6 is repetitive work that software handles well, while the judgment in steps 1 and 3 stays with the investor. There are two ways to set it up with Sonexa Labs.
Juno is our ready-to-use deal sourcing software. It helps investors discover emerging companies through early building activity, investigate founder and company context, keep companies on a radar and track outreach.
When a fund's sourcing logic, sources or outputs need to be its own, we build a fund-specific discovery system around the firm's mandate and review process instead.
Either way, the decision about which founders to back stays with your team.