The numbers in venture capital are unforgiving. A typical early-stage fund sees 1,000 to 3,000 deals per year and invests in 8 to 15. That means you need a system that can efficiently process 99% of your deal flow while ensuring the top 1% gets the attention it deserves.
Top-performing funds do not achieve this through superhuman pattern recognition or working 16-hour days. They achieve it through deliberate sourcing strategies, structured screening workflows, and disciplined time allocation. The best funds are not just better at picking winners. They are better at seeing winners in the first place and moving quickly when they do.
This post breaks down how high-performing VC funds actually source and screen at scale.
The Sourcing Mix: Where Deals Really Come From
Every fund talks about their proprietary deal flow. In practice, most funds source from the same five channels, just in different proportions:
1. Warm Referrals (40 to 60% of pipeline)
This is the dominant sourcing channel for nearly every established fund. Warm referrals come from:
Portfolio founders: Your existing founders know the ecosystem. They see what other founders are building, who is hiring well, and who has real traction. Portfolio founder referrals tend to be high quality because the referring founder has reputational skin in the game.
Co-investors and other GPs: The venture ecosystem is deeply networked. A growth-stage fund sees a company too early for them and passes it to you. A seed fund in another geography sees a company expanding into your market. These cross-fund referrals are reliable deal flow when you cultivate the relationships.
Angels and scouts: Many funds formalize this channel through scout programs. Pay experienced operators or angels a small carry allocation to source deals. The best scout programs generate 20 to 40 high-quality referrals per year per scout.
Advisors, lawyers, and accountants: Startup lawyers and accountants see companies before most VCs do. They know who is incorporating, who is raising, and who has interesting traction. These referrals are often early, which is valuable if you invest at pre-seed or seed.
Why this matters: Referral quality is measurable. Track conversion rates by referral source over time. You will find that a small number of sources (perhaps 10 to 20 individuals) generate a disproportionate share of your invested deals. Once you identify these high-signal sources, invest heavily in those relationships.
2. Cold Inbound (20 to 30% of pipeline)
Cold inbound is the pitch deck that arrives unsolicited via your website contact form, your general email address, or a LinkedIn message. The reputation of cold inbound is poor, and the conversion rate is genuinely lower than warm referrals. But writing it off entirely is a mistake.
Several billion-dollar companies were initially cold inbound to their lead investors. The signal-to-noise ratio is low, but the signal is real.
The key to managing cold inbound is not spending more time on it. It is spending time more efficiently:
Standardize the intake: If every cold inbound arrives in a different format (some send decks, some send one-pagers, some send a three-paragraph email), screening becomes inconsistent. Create a standard submission process where possible, or parse submissions into a consistent format automatically.
First-pass triage: Spend no more than 2 to 3 minutes on initial triage. Does this match our thesis? Is there a team signal worth exploring? Is there any traction data? Yes, no, or maybe. That is the only decision at this stage.
Batch processing: Handle cold inbound in batches, not one at a time as they arrive. Dedicate 30 to 60 minutes per day (or have an associate do it) to process the inbox. This is far more efficient than context-switching throughout the day.
3. Events and Conferences (5 to 10% of pipeline)
Demo days, industry conferences, pitch competitions, and founder meetups. The sourcing value of events is often overstated by the events themselves and understated by funds that attend them grudgingly.
The truth is nuanced. Events are valuable for two things:
Direct sourcing: You see 20 companies pitch at a demo day. Two are interesting. You schedule follow-ups. This works, but the hit rate is similar to cold inbound.
Relationship building: The real value of events is meeting the referral sources who will send you deals for years. The angel you sit next to at dinner. The founder who just exited and is thinking about investing. The corporate development person who sees interesting startups through partnership discussions.
Top funds track which events produce deal flow (directly or through relationships formed) and ruthlessly cut the ones that do not. Three or four high-value events per year is usually more productive than attending 20.
4. Outbound Research (5 to 15% of pipeline)
Proactive sourcing, where you identify companies before they come to you. This can mean:
Market mapping: Systematically identifying every company in a space you find interesting. If you have a thesis on vertical SaaS for construction, find every company building software for contractors, general contractors, and construction project managers.
Talent tracking: Following where strong operators go after they leave successful companies. When three ex-Stripe engineers start something together, that is a signal worth pursuing.
Data-driven sourcing: Using product analytics (website traffic growth), hiring data (who is scaling their engineering team), or app store rankings to identify companies with traction before they start fundraising.
Outbound sourcing is time-intensive but can produce differentiated deal flow. If you reach a founder before they have 30 term sheets, you have a meaningful advantage in winning the deal.
5. Accelerator and Incubator Networks (5 to 10% of pipeline)
Y Combinator, Techstars, 500 Global, and similar programs are concentrated sources of deal flow. Demo day is the obvious access point, but the best engagement happens earlier, through mentor relationships, office hours, or batch interactions.
The challenge with accelerator sourcing is competition. Every VC sees the same demo day. The advantage goes to funds with relationships inside the programs, giving them access before demo day, and funds that can evaluate and move quickly in the compressed fundraising window.
The Screening Workflow: From 1,000 to 100
If sourcing is about volume, screening is about velocity. You need to reduce 1,000 deals to 100 meetings without spending 30 minutes on each one.
The Two-Tier Screening Model
Most high-performing funds use a two-tier screening process:
Tier 1: Associate/Analyst Screen (2 to 5 minutes per deal)
This is a quick filter applied to every deal that enters the pipeline. The screener is looking for immediate disqualifiers:
- Outside fund mandate (wrong stage, wrong sector, wrong geography)
- Team does not meet minimum bar (no relevant experience, first-time solo founder in a complex domain)
- Market too small for venture-scale returns
- No evidence of any traction or insight that differentiates the company
Tier 1 screening should eliminate 50 to 70% of deals. The decision output is simple: pass, advance to Tier 2, or flag for immediate partner attention (high-priority referral, repeat founder, etc.).
Tier 2: Principal/Partner Screen (10 to 15 minutes per deal)
Deals that survive Tier 1 get a deeper look. This typically involves:
- Reading the full pitch deck (not just skimming)
- Reviewing the team's background in more detail
- Checking for any existing portfolio conflicts or co-investment opportunities
- Comparing to other companies in the pipeline in the same space
- Making the meeting/no-meeting decision
Tier 2 screening should narrow the funnel to the 10 to 15% of deals that warrant a meeting.
Screening Criteria That Actually Work
Every fund's screening criteria are different, but the structure is usually similar. Here are the dimensions that consistently matter:
Team (highest weight at early stage)
- Founder-market fit: Does this team have a unique right to win in this market?
- Previous startup experience (not required, but a positive signal)
- Technical depth appropriate for the product
- Ability to recruit strong early employees
Market
- Total addressable market large enough for venture-scale outcome
- Market growing or undergoing structural change
- Timing: why is now the right time for this product?
Product and Traction
- For pre-product companies: clarity of vision and evidence of customer discovery
- For post-launch companies: growth trajectory, retention, and unit economics
- Differentiation from existing solutions
Fundraise and Deal Terms
- Raise amount appropriate for stage and milestones
- Valuation in range for your fund's portfolio construction
- Investor syndicate quality (if round is partially filled)
- Pro rata rights and governance terms
How to Say No Fast
The ability to say no quickly and gracefully is one of the most important operational skills in venture capital. Every minute spent deliberating on a clear "no" is a minute not spent on a potential "yes."
Build a list of automatic pass criteria. These will vary by fund, but common ones include:
- Company is raising at a stage outside your mandate
- Sector is explicitly outside your thesis
- Geographic focus does not match your fund
- Founder has disclosed legal or ethical issues
- Company has raised from investors your fund has policy conflicts with
For these cases, do not deliberate. Pass immediately with a respectful, templated response. Speed benefits everyone: the founder can focus on investors who are a fit, and you can focus on deals that match your criteria.
For less clear-cut passes, develop a decision framework with a time limit. If you cannot decide in 15 minutes whether a deal deserves a meeting, the answer is usually no. True conviction rarely requires extended deliberation at the screening stage.
The Feedback Loop: Learning from Passes
Your passes are as informative as your investments, possibly more so. Periodically review companies you passed on:
- Which passed companies raised successful subsequent rounds?
- Which passed companies achieved significant traction?
- Were your pass reasons validated or invalidated?
This analysis reveals systematic blind spots in your screening. Maybe you consistently underweight certain founder profiles. Maybe your market size estimates are too conservative in specific sectors. Maybe deals from certain sources deserved more attention than they got.
One practical approach: every quarter, pull up the 20 most notable companies you passed on in the prior year. Categorize the pass reasons and evaluate whether the pass was correct with the benefit of hindsight. This is not about self-flagellation. It is about calibrating your judgment.
The Associate vs. Partner Division of Labor
In funds with both associates and partners, the division of labor in sourcing and screening matters enormously.
Associates should own:
- First-pass screening of all inbound deal flow
- CRM data hygiene and pipeline management
- Market mapping and outbound research
- Initial relationship building with founders (coffee chats, ecosystem presence)
- Preparing screening memos for partner review
Partners should own:
- Tier 2 screening decisions (meeting/no-meeting)
- All first meetings with founders
- Referral source relationship management
- IC presentations and investment decisions
- LP-facing pipeline discussions
The most common failure mode is partners doing associate work (reading every cold inbound email) or associates doing partner work (making pass/meet decisions without guidance). Clear role definition and trust in the screening process are essential.
Building the System
None of this works without tooling. The operational backbone of a high-volume sourcing and screening workflow needs:
- Centralized deal capture: Every deal from every source lands in one system
- Structured pipeline stages: Clear stages with defined entry and exit criteria
- Source attribution: Know where every deal came from
- Collaboration features: Multiple team members working on the same pipeline
- Search and filter: Find any deal by any dimension instantly
- Automation: Email capture, pitch deck parsing, auto-tagging, and routing
Roulette handles these requirements for VC teams. It is designed specifically for the venture capital workflow, from initial pitch deck capture through portfolio management, with the automation and collaboration features that make high-volume deal processing manageable for small teams.
The fund that sees 1,000 deals and the fund that sees 3,000 deals have the same number of hours in a day. The difference is how they spend those hours. With the right sourcing strategy and screening workflow, your team spends time on judgment and relationship-building rather than data entry and email management.
That is where returns come from.