Field guide · Proprietary search

How to turn an acquisition thesis into an evidence-backed target pipeline

The hardest part of proprietary deal sourcing is not producing a long company list. It is applying the same acquisition thesis across a market, showing why each company belongs, and keeping the evidence current enough that someone can act on it.

A list is a snapshot. A target pipeline is an operating system.

This guide lays out a practical version of that system for independent sponsors, search funders, family offices, and lean private-equity teams. It is deliberately tool-agnostic. You can use it to audit a spreadsheet, brief an analyst, or decide what an automated research workflow must do before it earns a place in your process.

Start with a thesis that can reject companies

Most acquisition theses read well in an investment memo and perform poorly as search instructions. “Founder-owned industrial services businesses in the Southeast” sounds specific until two analysts interpret every word differently.

A working thesis needs explicit fields:

• Industry boundaries: the activities that count, the adjacent activities that do not,

and the evidence that distinguishes them.

• Geography: headquarters, operating footprint, customer concentration, or some

combination of the three.

• Scale: revenue, employees, locations, transaction value, or another observable proxy,

with a rule for missing data.

• Ownership: the forms of ownership you will include and the evidence required before

calling a company independent, founder-owned, family-owned, sponsor-backed, or public.

• Operating fit: the capabilities, customer mix, certifications, assets, or recurring

work that make the company relevant.

• Exclusions: business models, end markets, conditions, and conflicts that should stop

the review.

The exclusion rules matter as much as the positive criteria. A thesis that cannot say “no” will produce a large universe and push the real work downstream.

A compact thesis contract

Write each criterion in four columns:

This contract does two useful things. It makes analysts consistent, and it exposes the criteria that cannot be observed reliably before outreach. Those unknowns become questions, not invented facts.

Build coverage before collecting names

A strong search begins with the shape of the market. The point is not to claim that a public dataset contains every acquisition target. The point is to create a denominator you can inspect.

For a U.S. search, County Business Patterns can help frame establishment counts by industry, geography, and employment size. The SBA table of size standards is useful for seeing how “small” changes by industry, even when its regulatory definition is not the same as your investment criterion. State licenses, procurement registrations, trade associations, certification directories, and local records can add categories that general company databases miss.

Create a coverage matrix before the first enrichment run:

1. Break the thesis into industry and geography cells.
2. Name at least one primary discovery source for each cell.
3. Name the source that can challenge or corroborate the primary source.
4. Record the source date, known omissions, and retrieval method.
5. Count discovered, reviewed, accepted, rejected, and unresolved companies by cell.

Now “coverage” has a meaning. If one county-industry cell contains no companies, you can distinguish a thin market from a failed connector or an overly narrow query.

Separate discovery, qualification, and contact readiness

These are different decisions and should never collapse into one score.

Discovery asks: does this entity exist, and is it potentially inside the thesis?

Qualification asks: what evidence supports or contradicts the thesis criteria?

Contact readiness asks: is there a verified person, role, and business channel that an approved outreach workflow may use?

Keeping the stages separate prevents a familiar failure. A data provider returns a good email address, so the company gets treated as a good target. Contact data is not investment evidence. It should not rescue a weak fit, and a missing email should not erase a strong company from research.

Use distinct states such as:

• discovered
• research needed
• evidence complete
• analyst review
• accepted target
• rejected with reason
• contact research needed
• outreach eligible
• suppressed

Every state needs an owner, an entry rule, and an exit rule. “Researching” is not a state if nobody can tell what is missing.

Use an evidence hierarchy

Not all sources deserve the same weight. Define the hierarchy before the team encounters a hard case.

One reasonable order is:

1. First-party and official records: company pages, regulatory filings, licenses,

government registrations, and signed transaction announcements.

1. High-quality independent records: reputable industry publications, local business

reporting, and established trade directories.

1. Structured aggregators: company and people databases with visible provenance and

freshness.

1. Search snippets and inferred attributes: useful for discovery, not for a final fact.

SEC EDGAR illustrates why the distinction matters. The SEC provides submissions and XBRL data through public APIs, and says those structures are updated as filings are disseminated. That can be strong evidence for a public filer or disclosed transaction. It does not prove ownership or financial facts for a private company that never appears in the filing.

For every material fact, store:

• the value;
• the source URL or record identifier;
• when the source was observed;
• the exact evidence excerpt or structured field;
• the analyst or system that made the interpretation;
• confidence and contradiction state; and
• when the fact should be checked again.

Preserve contradictions. Two sources disagreeing is an operational fact, not a reason to silently select the convenient answer.

Score fit without hiding uncertainty

A ranking helps allocate analyst time. It should not pretend that incomplete evidence is complete.

Separate at least three concepts:

• Fit: how closely the known facts match the thesis.
• Evidence quality: how reliable and current the supporting sources are.
• Readiness: whether the record has passed the reviews needed for the next action.

An 88 fit score with weak ownership evidence is not more outreach-ready than a 76 with complete, current evidence. Show the missing criterion beside the score.

A simple scorecard can use weighted thesis criteria, but add hard stops:

• A confirmed exclusion rejects the company regardless of the total.
• A required unknown holds the company for review.
• Stale evidence cannot authorize contact.
• A suppression record overrides every commercial score.

This is less impressive than a single magic number. It is much easier to trust.

Make the output answer the next decision

The right unit of output is not a profile. It is a decision-ready target record.

For each company, the reviewer should be able to see:

• the one-line fit verdict;
• the criteria that passed, failed, or remain unknown;
• the evidence beside each material assertion;
• the reason it outranks the next company;
• the exact review or research step still required;
• ownership and contact readiness as separate sections; and
• the history of prior decisions.

The reviewer should be able to reject a company with a durable reason, request one specific fact, or approve a handoff without rebuilding the research.

Run a weekly control loop

Proprietary sourcing becomes durable when the search keeps learning without quietly changing the thesis.

Review these questions every week:

1. Which market cells produced accepted targets, and which produced only noise?
2. Which criterion created the most unresolved records?
3. Which sources changed, failed, or became stale?
4. Which rejection reasons repeat often enough to become a formal exclusion?
5. Did outreach reveal a thesis assumption that research could not observe?
6. Which accepted targets moved into a substantive conversation?
7. What changed in the thesis, who approved it, and which records must be re-evaluated?

Do not optimize to the size of the list. Optimize to decision-ready coverage and qualified conversations.

A worksheet for your next mandate

Before adding another data source, answer these ten prompts:

1. What exact acquisition outcome is this mandate pursuing?
2. What observable facts define a pass?
3. What confirmed fact creates an immediate rejection?
4. Which required facts are routinely unavailable?
5. What is the market-coverage denominator?
6. Which source is authoritative for each material fact?
7. How long may each fact remain unverified?
8. What must a reviewer approve before contact research begins?
9. What must be true before outreach begins?
10. Which commercial outcome will tell you the system is improving?

If the answers are unclear, a larger database will give you more ambiguity faster.

If the answers are clear, the work can become repeatable. DealPort Workbench is built for that second case: recurring analyst capacity configured around the way a lean acquisition team defines, researches, reviews, and advances a target. Target Research is one Solution archetype inside that Workbench, not a separate product or a promise of a deal.

Bring one live or planned acquisition thesis to a review. We will map the criteria, evidence, decision points, and handoff together. You can judge the result against a market your team already knows.

Sources and further reading

• DealPort platform, observed September 6, 2026.
• SEC EDGAR application programming interfaces, observed September 6, 2026.
• U.S. Census Bureau County Business Patterns, observed September 6, 2026.
• U.S. Small Business Administration table of size standards, observed September 6, 2026.