Building a Repeatable Deal-Sourcing Process for a Small Fund

Large funds can afford a team of associates whose job is to find companies. Most small funds, family offices and angel groups cannot. Their sourcing depends on whoever happens to forward an introduction, which means the pipeline is only as good as the partners' inboxes.

A repeatable process changes that. It does not need to be complicated. It needs to be written down, run every week, and shared by everyone who makes decisions.

Start With a Thesis You Can Screen Against

"We invest in great founders" cannot be screened. "Pre-seed and seed B2B software for logistics companies in Southeast Asia, with early revenue or a pilot customer" can. A thesis written that precisely tells you which signals matter, which regions to watch, and which companies to drop quickly. We wrote a separate guide on writing an investment thesis you can screen against.

Pick Your Sources Deliberately

List the places where companies that fit your thesis appear first. For most funds, that means a mix of:

• accelerator batches and early-stage fund portfolios in your regions;
• startup press in the local languages of those regions;
• launch platforms and product directories for your sectors;
• open-source and AI registries if you invest in developer tools;
• company registries and private-placement filings where they are public.

Write the list down. A source that is not on the list is not part of the process.

A Weekly Rhythm

1. Daily, ten minutes: scan new companies that match your saved searches. Save anything interesting to a shared shortlist with one line on why.
2. Weekly, one hour: review the shortlist as a team. For each company, decide: reach out, keep watching, or pass. Write down the reason.
3. Monthly: look at what you passed on and what happened to it. This is how your filters get better.

Triage Rules That Save Hours

Most of the time in sourcing is spent deciding what not to look at. A few written rules make that fast and consistent across the team:

• Out of thesis, out of the queue. If a company fails on stage, geography or sector, pass immediately and do not revisit unless something material changes.
• One kind of evidence is a "watch", not a "meet". A single launch or a single directory listing earns a place on the watch list, not a call.
• Two independent signals in a month earn a closer look. An accelerator batch plus a first hire, or a launch plus open-source traction, is worth thirty minutes of research.
• Dead websites end the review. If the site is parked or offline, move on.

Outreach That Gets Replies

Founders get many cold messages. The ones that get answered share three traits: they are short, they show you did your homework, and they ask for something small.

Saw your launch last week and the new hire for a second engineer. We invest at pre-seed in tools for freight brokers. Would a 20-minute call in the next two weeks be useful?

Mention the specific signals that caught your attention. It tells the founder you found them for a reason, and it makes it easy to say yes.

Keep One Pipeline

The most common failure is not missing companies. It is losing them. A founder mentioned in a call, a spreadsheet only one partner uses, a note in someone's email: all of it disappears. Keep one pipeline, with simple stages everyone understands: watching, contacted, meeting, diligence, invested, passed. Every company lives in exactly one stage.

Measure the Process

You do not need a dashboard. Track three numbers each month: how many new companies entered the pipeline, how many you met, and how early you met them compared with their first public round. If the third number improves, the process is working.

AlphaScout was designed around this routine: saved searches with daily or weekly alerts, a pipeline board shared with your team, and theses that are scored against every new company each morning. Browse last week's hidden gems for free, then set it up in an afternoon.