The Target-Company List Standard: When Your Source List Is Ready to Work

You have a spreadsheet of companies you plan to pull candidates from. This standard tells you whether that list is complete, balanced, and safe enough to start working, or whether you are about to burn two weeks sourcing from the wrong places. It is written for in-house recruiters, sourcers, talent leaders, and founders doing their own hiring, and it delivers six pass/fail criteria that two people apply to the same list and reach the same verdict.

Every published resource explains how to build a target-company list. Almost none define when the list is finished. So sourcers guess at coverage, tier balance, and freshness, then discover two weeks in that the list was skewed, stale, or thin. This turns those guesses into a gate you run before any names get pulled.

What "ready to work" means for a target-company list

A target-company list is ready to work when it passes six criteria: scoped, categorized, tiered on one axis, coverage-measured, fresh, and lawful. If any one fails, the list is not-yet, no matter how many companies it holds.

The target company universe is the pool from which you identify passive candidates, and it is the foundation of all market mapping. That makes its quality upstream of everything downstream. A skewed universe produces a skewed shortlist; a stale one produces bounced outreach; an unlawful one produces a fine. The point of a standard is that "good enough to work" stops being a feeling and becomes a check two people grade the same way.

The six criteria are deliberately narrow. Count is not one of them, because count is the thing everyone already tracks and the thing that most often lies. A 40-company list can fail balance, coverage, freshness, and lawful basis all at once while looking finished. Grade the criteria, not the row total.

Criterion 1: Count is scoped, not padded

The count is right when the list holds 20 to 50 target companies for a talent map, or 10 to 30 for a single-role sourcing ICP, and every company earns its place against a written scope. Count alone proves nothing; it is a floor, not a verdict.

Practitioner talent maps converge on a tight range. Most maps focus on 20 to 50 target companies rather than the full market. Sourcing ICPs, which are the profile of the ideal prospect, run tighter, listing 10 to 30 companies where those prospects currently work. Job-search worksheets tier even smaller, putting 4 to 6 companies in active pursuit at the top.

Use the range that matches your job. A whole-function talent map runs wider than a single-req ICP. What the count must clear is the scope test: could a second sourcer read your written scope and predict roughly which companies belong? If the scope is "senior backend engineers, Series B fintech, US" then a company should be defensible against that sentence or removed. A vague scope produces a spreadsheet nobody uses.

For sizing effort, hold this benchmark: a focused map covering 20 to 30 target companies and 100 to 200 named candidates typically takes a sourcer two to four weeks full-time, depending on market depth and data quality. Broader maps take longer. Tightly scoped ones can be done in a week. If your list is far outside 20 to 50 and you cannot say why, that is a signal the scope is doing no filtering.

Criterion 2: Balance across the three required categories

The list is balanced when every company carries exactly one category tag - direct competitor, adjacent industry, or feeder organisation - and competitors do not crowd out the other two. The documented starting ratio is ten direct competitors to five adjacent industries.

Three categories recur across every credible source:

• Direct competitors. Companies hiring for the same roles you are. The obvious pool, and the one everyone over-weights.
• Adjacent industries. Companies where similar skills are developing under a different label. Adjacent industries are where the under-priced talent often hides, which makes adjacency a pricing arbitrage, not a nicety.
• Feeder organisations. Academy companies that reliably develop great talent, plus your own past pipeline. This is where your ATS lives.

The 10:5 competitor-to-adjacent ratio is the only published proportion, so treat it as a starting point rather than a hard law. Beyond it, precise proportions are not established publicly. What is gradeable is the presence of all three categories and the absence of a competitor monoculture. If adjacents and feeders together are a rounding error, the list forfeits the cheapest talent and the highest-converting source in one move.

That highest-converting source is your own database. In 2024, 44 percent of sourced hires came from people already in a company's CRM or ATS, up from 29.1 percent in 2021. A list that ignores warm ATS companies leaves the single best source untagged. Tag them as feeders and the list gets both more complete and more likely to convert.

Criterion 3: One tier axis, applied consistently

The list is tiered when each company sits in a tier assigned by a single stated criterion, and that criterion is written at the top of the sheet. Mixing two tiering axes in one Tier 1 is a documented failure mode.

Two conventions dominate, and they measure different things:

The skill-based scheme reads: Tier 1 is a strong skill match at a competitor with signals of openness to move; Tier 2 is solid skills that need some growth or role adjustment; Tier 3 is a watch list to monitor over time. The company-fit scheme sorts companies A/B/C on client fit and distinguishes client, prospect, and not-a-fit.

Both work. What breaks is using both at once, so that Tier 1 means "great candidates" for half the sheet and "priority accounts" for the other half. Two graders will then disagree on where a company belongs, and the standard collapses. Pick the axis that matches your job - skill match for a specialist search, company priority for account-based sourcing - and write it down.

Criterion 4: Coverage measured against a stated segment

Coverage passes when you can state what share of a defined segment the list captures, and that share is 80 to 90 percent. The 80 to 90 percent target is meaningless without a denominator you can name.

The goal of a map is breadth: most maps aim for 80 to 90 percent coverage of the defined segment. This is the only published completeness percentage, which makes it load-bearing and single-sourced. The failure it guards against is a percentage asserted with no denominator. "We cover 85 percent" is not a claim until you say 85 percent of what: how many companies, or how many candidates, define the full segment.

Two things silently cap coverage, and both are visible in supply data. The first is skill scarcity. The second is geography.

In Refolk's index of professional profiles, US software engineers listing Go number 2,767 against 607 listing Rust. A Rust list will exhaust roughly 4.6x faster than a Go list at equal outreach volume. So "80 to 90 percent coverage" of a scarce-skill segment is a far smaller absolute pool than the same percentage of an abundant one, and your outreach math has to account for it before you commit send volume.

Geography caps completeness the same way. A playbook proven in the US will hit a wall where the same segment is an order of magnitude smaller.

The UK pool for this segment is 3.4 percent of the US pool, roughly 29 times smaller. Coverage targets must be market-specific, not global. A list that hits 85 percent in a deep market and 85 percent in a thin one is describing two very different amounts of workable supply.

Coverage is a fraction, and a fraction is worthless until you can name its denominator out loud.

This is the point where measuring the universe by hand starts to cost more than the list is worth. Refolk lets you describe the segment in plain English - competitors, adjacents, and geography in one query - and returns the people grouped by employer, so you can size each tier against real supply instead of guessing the denominator.

Criterion 5: Freshness verified with a date

Freshness passes when the list carries a last-refresh date inside one quarter for high-velocity roles like engineering and sales. A list without a dated refresh is failing this criterion even if count, balance, and coverage all pass.

Maps decay fast. People change jobs, get promoted, leave the market. Most mature TA teams refresh strategic maps quarterly for critical roles. Freshness is the axis process guides omit, which is exactly why it belongs in the gate: a beautifully balanced list of companies whose teams have since scattered is worse than no list, because it looks trustworthy while being wrong.

Cadence varies by role velocity:

• High-velocity functions - engineering, sales - refresh quarterly.
• Lower-velocity roles - executive, specialist - can run annually, with a light pass when a req opens.
• Trigger-based refresh - re-run when compensation benchmarks shift or competitor activity creates a talent-supply event.

A quantified decay rate per month is not established publicly, so do not put a false precision on it. The gradeable version is simple: is there a date, and is it recent enough for this role's velocity? An undated list fails. A list last touched eight months ago for an engineering req fails. The date is the evidence; the absence of a date is itself a failure.

Criterion 6: Lawful basis on file before names get pulled

Lawful basis passes when a dated legitimate-interest assessment is attached for EU/UK candidates, or a documented antitrust check for the US, and no reciprocal no-poach understanding underlies any company's inclusion or exclusion. Legality is decided at list-construction time, not at outreach.

For EU and UK candidates, the documented basis is legitimate interest, not consent. Legitimate interest is the most common basis for active recruitment. It requires the ICO's three-part balancing test: the interest must be real and present, the processing must be necessary, and the candidate's rights must not override your interest. Transparency has a clock too - you must tell sourced candidates you are processing their data within one month of first processing it. "It's a public LinkedIn profile" is not a lawful basis; a documented assessment is.

This is not theoretical. Regulators have issued EUR 6.8 billion in GDPR fines since 2018, and the employment sector alone has taken 193 fines. The Irish regulator fined LinkedIn EUR 310 million over invalid consent and legitimate-interest claims. The list is the compliance surface.

In the US, the exposure is at the list level and it is antitrust. No-poach and wage-fixing agreements between businesses that compete for employees can violate antitrust law even if they cause no actual harm. Crucially, the agreement does not need to be written: mutual understandings or gentlemen's agreements can run afoul of the law. So if a company is on or off your list because of a "we won't poach each other" conversation, the universe is tainted at its source.

The procedure: from raw pool to graded list

Run these seven steps in order. Steps one through six build the list; step seven grades it. The whole sequence for a 20 to 30 company map runs two to four weeks full-time, with the grading step taking under an hour once the six criteria are in place.

The two-grader step is not ceremony. It is how you find out whether your criteria are actually gradeable. If two sourcers disagree on a verdict, the disagreement points straight at the underspecified criterion - usually the tier axis or the coverage denominator - and you tighten it before the list is worked.

How this goes wrong: failure modes and false positives

Most lists fail the gate for one of eight reasons, and each has a false positive that makes a broken list feel finished. This is the section to read twice, because a standard that overclaims is worse than none.

• Count looks done but is skewed. A 40-company list that is 38 direct competitors fails balance. The false positive is a high count masking zero feeder or adjacent coverage. Check: a category tag on every row and a competitor-to-adjacent ratio near 10:5.
• Coverage claimed, not measured. "80 to 90 percent" asserted without a denominator. The false positive is that a big list feels complete. Check: state the segment size the percentage is computed against.
• Stale list treated as fresh. Named companies whose teams have since scattered. Check: a last-refresh date within one quarter for engineering and sales.
• Tier scheme conflates two axes. Mixing "strong skill match" with "company fit" in one Tier 1. Check: one stated tier criterion applied consistently by both graders.
• Thin-supply market read as workable. A Rust list of 607 worked like a Go list of 2,767 burns out fast. Check: pool size per tier before committing outreach volume.
• No lawful basis on file. EU/UK sourcing without a documented LIA, with "public profile" assumed sufficient. Check: the real-and-present, necessary, rights-not-overridden test is written down and dated.
• List built via a competitor conversation. A "we won't poach each other" understanding taints the universe. Check: no reciprocal restraint underlies any inclusion or exclusion.
• Tool mistaken for the discipline. The mapping discipline is the analytical layer that decides what data goes in, how often it refreshes, and which segments matter. Confusing the tool for the discipline is one of the most common reasons these efforts stall. Check: the list has stated coverage, balance, and refresh rules, not just a CRM export.

The through-line is that every false positive is something looking finished. The count is high. The percentage is stated. The CRM exported cleanly. The gate exists precisely to distrust appearance and demand evidence: a tag, a denominator, a date, an assessment.

The ready-to-work checklist

Run this before any names get pulled. Every item must pass. One failure means not-yet.

Keeping the standard current

A ready-to-work list stops being ready the moment the market moves, so adopt the gate as a recurring policy, not a one-time pass. Re-run the full grade on the refresh cadence the roles demand, and re-run the freshness and lawful-basis criteria on triggers even between scheduled refreshes.

Set the cadence by role velocity: quarterly for engineering and sales, annually with a light pass at req-open for executive and specialist roles. Add trigger-based re-grades when a competitor has a layoff, a compensation benchmark shifts, or a talent-supply event opens a window. On each re-grade, the two criteria most likely to have decayed are freshness, which is time-bound by definition, and lawful basis, which changes if your candidate geography or processing changes.

Treat the two-grader agreement as your calibration check. When two sourcers stop reaching the same verdict, the market has drifted past your written criteria, and the fix is to re-tighten the scope, the tier axis, or the denominator - not to override the disagreement. Balance beats size, freshness beats both, and the gate is only worth running if it produces the same grade in two hands.