Step 1. Define your thesis
Every effective search starts narrower than feels comfortable. A thesis is a specific statement of what you are looking for: an industry or set of adjacent industries, a geography you can realistically operate in, and a size range, usually expressed in employee count or estimated revenue. Without this, sourcing becomes an unfocused browse through millions of off-market businesses with no way to prioritize any of them.
A workable thesis sounds like: commercial services companies, Western Canada, 15 to 75 employees, recurring revenue model. It is specific enough to filter a database meaningfully and broad enough to leave you a real universe to work with, typically several thousand to tens of thousands of companies.
Step 2. Build the universe
Once the thesis is set, the next job is assembling every company that could plausibly fit it, not just the ones already for sale. This is where public data does most of the work: business registrations, licensing records, and government program data all contain company names, addresses, industry codes, and sometimes employee or payroll figures. The Paycheck Protection Program dataset, released publicly by the SBA, is a particularly rich free source for this, since it names borrowers by NAICS code and location. See the PPP loan company lookup page for how that dataset works in detail.
The goal at this stage is coverage, not precision. A universe of 10,000 companies that roughly fits the thesis is more useful than a hand-picked list of 40, because the qualifying and filtering happens in the next step.
Step 3. Qualify with estimates
Almost none of these companies publish real financials. Qualifying means using what is observable, employee count, years in operation, number of locations, industry benchmarks, to produce an estimated range for revenue or seller discretionary earnings (SDE). A responsible estimate is shown as a range, not a point figure, and is labeled with a confidence level: high, moderate, or low, depending on how much direct evidence supports it.
This is also the step where a buyer separates a shortlist worth investigating from the long tail that does not fit. A company estimated at 4 to 7 million in revenue with moderate confidence is worth a closer look; one estimated at under 500,000 with low confidence probably is not, for a thesis targeting mid-sized operating companies.
Step 4. Track and watch
A shortlist is only useful if you keep it organized and keep watching it for change. Save the universe you built, create lists for different tiers of priority, and set watches for signals: a leadership change, a new location, a shift in hiring activity, or a change in licensing status. None of these signals confirm a company is for sale. They are simply reasons to move a company up your list and take a closer look before reaching out.
Step 5. Reach out and verify
A first message should be short, specific, and honest about who you are and why you are writing. Reference something real about the business, ask an open question about whether the owner has ever considered a transition, and avoid anything that reads as a mass-produced form letter. If a broker is involved instead of the owner directly, basic etiquette applies: be clear about your financing status and timeline before asking for financial detail.
Once a conversation starts, treat every estimate you built in step 3 as a hypothesis to be tested, not a fact. Real financials, tax returns, and management accounts replace estimated ranges the moment an owner shares them. The estimate’s job was only ever to help you prioritize which companies were worth this conversation in the first place. For a broader view of how this fits into a professional sourcing program, see search fund deal sourcing, or return to SIFT home to start building a universe against your own thesis.
Frequently asked questions
How long does off-market sourcing usually take?
Most buyers spend two to six months moving from a defined thesis to a first serious conversation, and longer to a signed letter of intent. That timeline depends heavily on how narrow the thesis is and how much time the buyer commits weekly to research and outreach. Treat any faster promise with skepticism; sourcing is a volume-and-patience process, not a shortcut.
What response rate should I expect from owner outreach?
Response rates on a well-targeted, personalized letter or email typically land in the low single digits to low teens as a percentage of contacts, and this varies by industry, letter quality, and how well the company fits the buyer's stated thesis. A one percent response rate against a list of a few hundred companies can still produce several real conversations.
Do I need a broker to find off-market businesses?
No, but a broker relationship is a useful complement, not a replacement. Brokers sometimes hear about a potential seller before any listing exists and can make an introduction with more built-in trust than a cold letter. Most disciplined searchers run direct outreach and broker relationships in parallel rather than choosing one.
What company data is available for free?
Public sources include state and provincial business registrations, professional licensing boards, and government program data such as the SBA Paycheck Protection Program dataset. SIFT combines sources like these into a single searchable universe of 3.85 million companies, growing toward roughly 10 million, free to search with CSV export up to 10,000 rows.