Deal Sourcing Infrastructure for Search Funds Under $10M

Traditional search fund playbooks assume access to intermediary networks, alumni databases, and proprietary deal flow from operating partners. If you're raising a $3M–$8M search fund targeting boring businesses in the lower middle market, none of that infrastructure exists. You build it yourself, or you spend 24 months cold-calling from a spreadsheet someone sold you on LinkedIn.

The infrastructure stack

Proprietary sourcing at search fund scale requires four layers: a target database, an enrichment layer, an outreach system, and a CRM that tracks every touchpoint. Most searchers have a CRM and nothing else. That's why pipeline feels like luck.

Layer 1: Target database

You need every operator in your thesis geography and vertical, not a sample, not a 'top 50 list.' For HVAC in Texas alone, that's 3,000–8,000 licensed contractors depending on how you define the market. Your database should include contact fields, digital presence signals, and a freshness timestamp.

Layer 2: Enrichment

Raw business names aren't leads. Enrichment adds: phone, email, website status, review velocity, owner neglect score, and ideally legal entity data. This is where off-market edge compounds, you're not calling the same list as every other searcher.

Layer 3: Outreach system

• Tier 1 (high neglect score, no website): personalized direct mail + phone follow-up within 7 days.
• Tier 2 (solid operator, aging digital presence): email + LinkedIn if owner is identifiable.
• Tier 3 (active digital presence): lower priority unless other signals emerge.

Layer 4: CRM discipline

Every contact gets: date first touched, channel, response status, next action date. Search funds that track this religiously convert 2–4% of outreach to meetings. Those that don't track convert under 0.5% and blame 'market conditions.'

Capacity planning for outreach

A two-person search team can personally work 300–500 active targets before CRM hygiene collapses. Beyond that, hire an analyst or narrow geography. Funds that import 10,000 rows and contact 200 leave false precision in LP decks, universe size matters less than meeting rate per Tier-1 target.

Operating cadence for a two-year search

Infrastructure only matters if it runs on a calendar. Weekly: refresh target universe in your thesis metros, re-score top 200 accounts, assign outreach quotas per searcher or analyst. Monthly: LP update with funnel metrics (contacted → meetings → LOIs → diligence). Quarterly: retire low-yield geographies and double list depth where response rate beats 2%. Search funds that treat sourcing as a project in month one and a habit in month eighteen outperform teams that binge-list-build then go quiet during diligence.

1. Monday: pipeline review, stale next-action dates, warm replies, broker leaks.
2. Tuesday–Thursday: outbound blocks (mail, phone, email) by tier.
3. Friday: log outcomes, update neglect tiers, prep one-page deal summaries for IC.

MCP and bounded research workflows

Sub-$10M funds cannot hire a data team. Agent-assisted research over a bounded API, city-level queries, no bulk export of the entire database, lets you answer IC questions fast: 'How many plumbing operators in Raleigh have no website and 50+ reviews?' That is the difference between a credible thesis slide and hand-waving about fragmentation.

What to avoid

• Buying bulk lead lists, they're sold to 50 other searchers simultaneously.
• Outsourcing sourcing entirely before you understand the trade.
• Spreading across 5 verticals before you've closed one deal in one vertical.

Handoff from search to close

When a target enters diligence, sourcing does not stop, it feeds comp sets and backup LOIs. Your database should tag every operator within 15 miles of the target for post-close tuck-in conversations. Roll-up-minded search funds win by showing LPs a platform story on deal two, not a one-off lucky close.

Team roles on a lean search

Traditional search pairs often split: CEO owns relationships and IC, CFO or associate owns model and lender process. On a three-person team, dedicate at least one FTE equivalent to list hygiene and outreach logging, otherwise your CRM becomes fiction. Interns can enrich; only trained searchers should run first owner calls on Tier-1 targets.

1. Define SLAs: returned owner calls within 4 business hours.
2. Weekly export of stale deals, no next action in 14 days gets killed or revived.
3. Document every broker introduction, avoid duplicate outreach embarrassing the fund.

Broker relationships without dependency

Search funds still take broker calls, but proprietary outreach should produce 60%+ of meetings in boring-business theses. Train searchers to ask brokers for off-market pocket listings, not recycled BizBuySell. When you bring a broker a buyer for their listing, ask for reciprocity: first look at their next HVAC listing in your geography.

Document broker-sourced vs. proprietary in CRM for LP reporting. Funds that lie about sourcing get caught in reference calls. Transparency about channel mix is fine if proprietary trend line improves quarter over quarter.

Security and data governance

Your target database is firm IP. Use role-based access in CRM, prohibit personal exports to Gmail drives, and train searchers on clean-desk policies for seller financials. LPs increasingly ask how you protect proprietary lists, a coherent answer is part of institutional fundraising.

Rotate access when analysts depart and audit export logs quarterly, one leaked target list teaches every owner in a trade to ignore search fund mail.