I tracked a firm's deal pipeline for an entire year. They reviewed 437 opportunities. Of those, 389 came from investment banks and brokers. Forty-one came from their network. Seven came from proprietary outbound origination. They closed two deals. Both were from the proprietary bucket.
That ratio isn't unusual. It's the dirty secret of PE deal sourcing. Brokered processes produce volume. Proprietary origination produces closings. Yet most firms allocate 90% of their sourcing effort to the channel that produces 10% of their results.
What M&A Deal Origination Actually Means
M&A deal origination is the process of identifying and initiating conversations with potential acquisition targets before those companies have engaged an investment bank or broker. It's the buy-side equivalent of business development, applied to finding companies to acquire rather than products to sell.
The distinction between origination and sourcing matters. Sourcing is passive: reviewing broker books, taking banker calls, monitoring auction calendars. Origination is active: mapping markets, identifying targets, building relationships, and creating opportunities that wouldn't exist without your effort. Preqin's 2025 survey found that 68% of PE firms rank deal sourcing as their most significant operational challenge. Most of them are using the wrong word. They mean deal origination.
Every closed PE transaction starts with origination, whether the firm recognizes it or not. Someone identified the company, someone made the first call, someone built enough trust to get the owner talking about selling. The question is whether that someone works for the PE firm or for a broker who's running a competitive process.
Primary vs. Intermediary-Sourced Deals
Two channels. Radically different economics.
Intermediary-sourced deals come through investment banks, M&A advisors, and business brokers. The sell-side advisor runs a structured process, contacts 50 to 200 potential buyers, and manages competitive bidding. These deals are, by definition, not proprietary. Every credible buyer in your sector sees them. According to Bain's 2025 Private Equity Report, average purchase multiples in competitive auctions run 11.2x EBITDA, compared to 8.7x for proprietary and lightly-brokered transactions. That 2.5x spread on a $10M EBITDA company is $25M in purchase price difference.
The auction dynamic is simple. More bidders means higher prices. Higher prices mean lower returns. Lower returns mean unhappy LPs. The intermediary model works well for sellers (which is why they hire bankers). It works poorly for buyers, who end up paying a market-clearing price that compresses IRR.
Primary-sourced deals come from the buyer's own origination efforts. Direct outreach to company owners, referral networks, industry conferences, and systematic market mapping. These conversations happen before a banker is involved, which changes the negotiation dynamic entirely. The owner isn't comparing 8 offers. They're having a conversation with one firm that's done its homework, understands their business, and showed up first.
Primary-sourced deals close at lower multiples, with less buyer competition, and often with better relationship dynamics through the diligence and transition period. The trade-off is volume. You'll review fewer primary-sourced opportunities than brokered ones. But your hit rate will be dramatically higher.
The M&A Origination Process
Five stages, each building on the last.
Stage 1: Thesis development. Before you can find targets, you need to know what you're looking for. Not "B2B software" but "vertical SaaS serving mid-market construction firms, $5M to $25M ARR, 70%+ gross margins, founder-owned, Southeast or Texas." Specificity is the engine of origination. Vague criteria produce vague results. The thesis should define industry verticals, geographic focus, revenue range, ownership type, and the strategic rationale for why you want to own these businesses.
Stage 2: Market mapping. Build the universe of companies that match your thesis. This isn't a PitchBook search. PitchBook captures roughly 40% to 60% of the lower-middle-market companies that meet typical PE criteria (their own coverage estimates acknowledge gaps in sub-$50M revenue companies). Real market mapping combines database research with primary-source data: state business filings, industry association directories, trade show exhibitor lists, permit databases, and proprietary web scraping of company information that hasn't been aggregated into commercial platforms.
A thorough market map for a focused search might identify 300 to 800 companies. That's the universe. Not the target list. The target list comes from the next stage.
Stage 3: Qualification and prioritization. Not every company in the universe is a real target. Qualification filters the universe based on transaction readiness signals: founder age (owners over 60 are statistically more likely to sell), recent equity partner changes, slowing growth, key customer losses, succession planning discussions, and technology disruption in their sector. AI tools can accelerate this stage dramatically. Natural language processing applied to news, job postings, and corporate filings can flag companies showing pre-transaction behavior patterns months before a formal process launches.
Stage 4: Outreach and relationship building. This is where most firms fail. They send a generic email that screams "private equity cold call" and wonder why response rates are 2%. Effective outreach is personalized, specific, and respectful. It references something concrete about the company. It explains why your firm is relevant to their situation. It doesn't ask "are you interested in selling?" on the first contact. (Nobody responds to that. Nobody.)
Multi-channel outreach works best. An introductory email, followed by a LinkedIn connection request, followed by a phone call, followed by a handwritten note if the target is high-priority. The cadence matters. The messaging matters more. Response rates for well-crafted, personalized origination campaigns run 8 to 15%, compared to 1 to 3% for generic outreach.
Stage 5: Pipeline management and conversion. An interested response isn't a deal. It's the beginning of a relationship that might take 6 to 24 months to produce a transaction. The best origination functions maintain relationships with hundreds of potential sellers simultaneously, staying in touch through quarterly check-ins, industry insight sharing, and genuine relationship building. The firms that treat origination as a one-touch process leave enormous value on the table.
AI's Role in M&A Deal Origination
AI is transforming deal origination in three specific ways. None of them involve replacing the deal team.
Target identification at scale. Machine learning models can process millions of data points (financial filings, web traffic, hiring activity, technology stack, news mentions) to score companies against a PE firm's investment criteria. What took an analyst a week to research manually, AI processes in minutes. Grata, for instance, now combines its platform with SourceScrub's data after their 2025 merger, creating a more comprehensive private company dataset than either had alone.
Timing intelligence. The most valuable insight in origination isn't which companies to target. It's when to call. AI models can identify pre-transaction signals (CEO stepping back from day-to-day operations, CFO hire at a previously founder-run company, new competitor entering the market, major customer contract expiring) that suggest a company may be approaching a transaction window. Reaching a founder six months before they call a banker is worth more than any database subscription.
Outreach optimization. AI can personalize outreach at scale, test messaging variations, optimize send timing, and predict which prospects are most likely to respond. But there's a ceiling. The most effective origination conversations are still human-to-human. AI gets you to the right door at the right time. A person has to walk through it.
Building vs. Buying Origination Capability
Three options. Each fits different firm profiles.
Build internally. Hire 1 to 3 dedicated business development professionals, invest in data infrastructure, and run origination as a core firm function. Cost: $300,000 to $750,000 annually (salaries, tools, data subscriptions, overhead). Time to full productivity: 9 to 12 months. Best for firms with $500M+ AUM, continuous deal activity, and the management bandwidth to oversee the function.
Buy externally. Engage a deal origination service to run outbound on your behalf. Cost: $60,000 to $300,000 annually. Time to first results: 60 to 90 days. Best for emerging managers, firms entering new sectors, and teams too lean to add dedicated origination headcount.
Hybrid model. Internal team handles relationship management and high-value targets. External partner handles market mapping, initial outreach, and lead qualification. This is where most sophisticated firms end up. The internal team focuses on conversion. The external partner focuses on coverage. Neither does the other's job well.
The worst option is the one most firms choose by default: assigning origination to junior deal team members as a side task. An associate who's running two active deals, supporting a portfolio company, and writing IC memos doesn't have the bandwidth to run systematic outbound origination. They'll do it poorly, get discouraged by low response rates, and default back to reviewing broker books.
Common Origination Mistakes
Confusing activity with results. Sending 1,000 emails isn't origination. Converting 15 of those into qualified conversations that lead to 3 management meetings is origination. Track conversion rates, not vanity metrics.
Starting too broad. "We're looking at all services businesses between $5M and $50M EBITDA" isn't a search. It's a haystack. Narrow the thesis, dominate a niche, then expand.
Ignoring the follow-up. 73% of proprietary deals close from the second or third contact, not the first. Most firms make one attempt and move on. The persistent ones win.
Underinvesting in data. If your market map is just a PitchBook export, you're seeing what everyone else sees. The firms winning proprietary deals invest in primary-source data that isn't available through any commercial platform.