How to Build a Deal Sourcing Pipeline from Scratch

Most firms don't have a deal sourcing pipeline. They have a collection of habits. Someone checks PitchBook on Monday, someone else talks to a banker on Thursday, and a few times a year the team goes to a conference and comes back with business cards. That's not a pipeline. A pipeline is a repeatable system that moves from thesis to target to conversation to tracking, and it runs whether or not anyone remembers to check a database.

Building one from scratch isn't complicated, but it does require being honest about where the gaps are.

What does a good deal sourcing pipeline look like?

A functional pipeline has six stages: thesis definition, target identification, initial screening, outreach, CRM tracking, and ongoing monitoring. Most firms skip the first stage and jump straight into searching, which is why results are scattered. If you can't describe your investment thesis in two sentences that a new analyst would understand, your sourcing will reflect that ambiguity. Every search, every conference conversation, and every banker relationship will drift slightly off target.

The stages don't need to be formal or heavy. They just need to exist. A thesis written on a whiteboard is better than one that lives only in a partner's head. A spreadsheet that tracks outreach is better than a CRM nobody uses. Start simple and add structure only when you feel the pain of not having it.

Why do most firms skip thesis definition?

Because it feels like busywork when you already know what you're looking for. But "we know it when we see it" is not a sourcing strategy. It's a screening strategy, and it only works when enough deal flow is already landing on your desk.

The numbers suggest it isn't. According to Sutton Place Strategies' Origination Benchmark Report, the median PE firm captures only 17.6% of relevant deal flow in its target markets. More than 80% of potential investments never reach the firm's desk. If your thesis is vague, that 17.6% is going to skew toward whatever your bankers happen to send you rather than what actually fits your strategy.

A good thesis for sourcing purposes is specific enough to generate a search but flexible enough to catch adjacent opportunities. "Lower middle market B2B services companies in the Southeast with $3-10M EBITDA" is a start. "Facility maintenance, janitorial, or landscaping businesses serving multi-site commercial customers, founder-owned, with recurring revenue above 60%" is something you can actually source against.

Where should deal flow come from?

Diversification matters in sourcing the same way it matters in a portfolio. If 80% of your deals come from one channel, you're exposed.

Intermediaries and brokers are the most common channel and the most competitive. Banker-led processes mean you're usually one of several bidders, and the information is symmetric. That said, strong banker relationships still matter. The best deals often go to the firm the banker calls first, before the formal process starts.

Proprietary databases like PitchBook, SourceScrub, and Axial give you structured data on a large number of companies. They're useful for filtering and screening but limited in that everyone has access to the same data. The edge comes from what you do with the results, not the results themselves.

Your network — portfolio company executives, advisors, accountants, attorneys, and other investors — is still the highest-converting source for most firms. Warm introductions convert at dramatically higher rates than cold outreach. The problem is that networks are hard to scale and tend to cluster around the same geographies and industries you already know.

Conferences and industry events are good for building relationships over time but poor for sourcing volume. They're better thought of as an outreach channel than a sourcing channel.

Agentic search tools are a newer category that fills a different gap. Instead of filtering on structured fields, tools like Radar let you describe what you're looking for in plain language and search across millions of companies using semantic matching. This is particularly useful for finding companies that fit your thesis but aren't in the databases yet — founder-led businesses that haven't raised capital, don't have a broker, and aren't actively marketing themselves. These are often the best targets and the hardest to find through traditional channels.

The right mix depends on your strategy. A lower middle market PE firm doing founder buyouts will lean more on networks and proprietary outreach. A growth equity firm looking at technology companies will lean more on databases and agentic tools. The point is to have multiple channels feeding the pipeline, not just one.

How should initial screening work?

Most firms over-screen and under-reach. They spend hours researching a company before making the first call, when that call would have answered their questions in five minutes. Initial screening should be fast and binary: does this company plausibly fit the thesis, and is there a reason to reach out?

At this stage you're looking for disqualifiers, not deep diligence. Wrong geography, wrong size, wrong business model, recent transaction, obvious distress you didn't want. If none of those apply, the company goes into outreach.

A common mistake is building elaborate scoring systems before you have enough data to calibrate them. Score companies after you've had 50 conversations, not before. Until then, your intuition about what matters is probably wrong in ways you won't discover until you're in the market.

What does outreach and tracking actually look like?

Outreach is where most pipelines die. Not because the firm can't find companies, but because nobody follows up. A target identified in January gets a first email in March and a second one never. The company sells to someone else in June.

The fix is simple but requires discipline. Every company that passes initial screening gets logged in a CRM or tracker with a next action and a date. The next action might be "send intro email," "ask portfolio CEO for warm intro," or "monitor for six months and revisit." The point is that no target disappears into a spreadsheet and stays there.

For outreach itself, keep it short and specific. Founders and business owners get dozens of generic PE emails. The ones that get responses reference something specific about the business — a recent hire, a new location, a product line that aligns with your portfolio. That level of specificity signals that you've done your homework, which is the minimum bar for a serious conversation.

Why is monitoring the most underrated part of the pipeline?

A one-time search is a snapshot. The market moves every day. Companies hit inflection points — a key hire, a new contract, a competitor exit — that change whether and when they're a fit for your thesis. The firms that show up first are the ones that have continuous visibility into their target market, not the ones that run a search quarterly and hope the timing works out.

Monitoring means watching for signals across your target universe: leadership changes, hiring patterns, funding events, geographic expansion, customer wins. Most of these signals are public but dispersed across LinkedIn, job boards, press releases, and company websites. Aggregating them manually is impractical at any real scale.

This is where agentic tools add the most value. Radar, for example, doesn't just find companies that match your thesis — it monitors them over time and surfaces changes that indicate a company might be approaching a transaction or an inflection point. That kind of continuous intelligence turns a static list into a living pipeline.

The firms that invest in monitoring end up with better timing on outreach, better context in conversations, and a compounding information advantage over competitors who are still working from point-in-time snapshots.

What's the simplest version of this that actually works?

If you're starting from nothing, here's the minimum viable pipeline:

1. Write down your thesis. Two to three sentences, specific enough that an analyst could source against it without asking you clarifying questions.
2. Pick two sourcing channels. One relationship-based, one technology-based. Run both for 90 days.
3. Screen fast. Spend no more than 10 minutes on initial screening per company. If you can't disqualify it quickly, reach out.
4. Log everything. Every company you look at, every outreach attempt, every response. A spreadsheet is fine to start.
5. Set up monitoring. Even a simple Google Alert on your target companies is better than nothing. A purpose-built tool is better than that.
6. Review monthly. Look at what's converting and what isn't. Adjust your thesis, your channels, and your outreach based on what you're learning.

That's it. You can build this in a week and refine it over the next quarter. The firms that source well aren't doing anything magical. They just have a system, and they run it consistently.

---

Radar helps PE, VC, and corp dev teams build sourcing pipelines that actually work. Describe your thesis in plain English, source across millions of private companies, and monitor your target market for changes — all in one place. Try it free or book a demo.