The Extraction-Friction Score: Rating a Candidate's Barriers to Moving

Before you commit weeks to a named passive candidate, decide how hard they will actually be to pry loose - and whether they are worth the effort now. This guide is for in-house recruiters, sourcers, talent leaders, and founders doing their own hiring. It gives you a single score, applied at sourcing time to one name at one employer, that turns "can I even move this person" into a written pursue-now, pursue-later, or deprioritize verdict before you write the first message.

Most of what is published on this covers one silo. A non-compete state tracker. An equity-handcuffs explainer. An H-1B transfer walkthrough. None of them combine into a pre-pursuit judgement you can apply to a specific person. That gap is expensive: the barriers you miss at sourcing time are the ones you discover at the offer stage, after the effort is already spent.

What the Extraction-Friction Score measures

The Extraction-Friction Score rates how hard a candidate will be to move across four independent barriers: contractual, immigration, financial, and geographic. Each is scored on its own, because each has a different owner, a different clock, and a different fix. A candidate can be legally free to leave and still financially chained; a candidate can be sitting on liquid equity and still stuck in a visa window.

The four dimensions are not equally common as hard stops, and treating them as equal is the first mistake. The contractual dimension gets outsized attention because it is loud and legalistic, but only about 18 to 20 percent of US workers are under a non-compete at all. Financial friction is quieter and far more often decisive, especially in states where non-competes are void.

The point of scoring before contact is triage. You have more names than hours. A candidate who scores low friction across all four is worth a message today. A candidate blocked on one clearable dimension - inside a visa window, or two months short of a vest cliff - is worth a calendared return, not a pursuit now. A candidate blocked on a dimension you cannot move is worth deprioritizing so your effort lands elsewhere.

The contractual dimension: does a non-compete actually bind here?

Start with the governing state, not the contract. Enforcement follows where the candidate physically works, not where they signed or what the choice-of-law clause says. This one fact flips more "no" verdicts to "yes" than any other, and it costs nothing to check.

US non-compete law splits three ways. First, a ban tier: California, Minnesota, North Dakota, and Oklahoma treat virtually all employee non-competes as void. California went further with SB 699 and AB 1076, effective January 1, 2024, voiding non-competes regardless of where they were signed and giving employees a private right of action with mandatory attorneys' fees. Second, a threshold tier: some states void non-competes below a wage line. Third, a reasonableness tier: Texas, Florida, and Georgia still enforce non-competes with reasonable scope, duration, and geographic limits, with courts applying the "blue pencil" doctrine to trim overbroad terms.

There is no federal backstop. The FTC finalized a rule in 2024 that would have broadly banned non-competes, a Texas federal court vacated it in August 2024, and as of 2026 no federal law categorically bans them.

The thresholds move annually, so re-check them rather than trusting a memorized number. Washington's employee line was $120,559.99 in 2024, $123,394.17 in 2025, and is cited at $126,858.83 for 2026. Washington also presumes any non-compete over 18 months unreasonable and carries a $5,000 statutory penalty for overbroad agreements. When you score a threshold state, compare the candidate's likely salary to the current line: below it, the restraint is void; above it, it may bind.

Now weight the score by role. Prevalence is not uniform, and assuming a non-compete exists because the employer is known for issuing them is a false positive that will cost you good candidates.

Senior and executive coverage runs roughly 3.4x to 3.7x the sub-$40k rate. The practical rule: weight the contractual dimension heavily only for senior, executive, physician, and trade-secret roles. For a mid-level individual contributor, treat the non-compete as noise unless you have direct evidence one exists and binds in their work state.

The immigration dimension: the 180-day clock that changes everything

For a green-card candidate, the load-bearing fact is 180 days of I-485 pendency under AC21. Once the I-485 has been pending 180 days, Section 204(j) portability lets the applicant change employers without restarting the green-card process, as long as the new role is in the same or similar occupational classification. Before that mark, the move is risky; after it, it is routine.

The danger window is real. If the change happens before 180 days, the original employer can withdraw the I-140 and jeopardize the pending I-485. After 180 days, an I-140 withdrawal generally does not invalidate the I-485. Timing, not persuasion, is what converts this candidate from risky to safe.

Three conditions must all hold for portability, and missing any one is a common scoring error:

• The I-485 has been pending at least 180 days.
• The I-140 is approved. The 180-day I-140 rule and the 180-day I-485 rule are independent, and both must be satisfied separately.
• The new role shares the same or similar SOC occupational classification, and Supplement J is filed.

For candidates earlier in the process, note the extension mechanics: AC21 Section 104(c) allows three-year H-1B extensions with an approved I-140 stuck in a backlog, and Section 106(a) allows one-year extensions once a PERM or I-140 has been pending 365 days. These tell you whether the candidate can keep working while their case matures, which affects how long a pursue-later window can safely run.

The scoring verdict for immigration is one of three states: portable (all three conditions met), inside the 180-day danger window (calendar the return), or pre-I-140 (deprioritize on this dimension unless the case is close and you can wait).

The financial dimension: the barrier that actually stops people

In ban states, financial friction is the more common hard stop than legal friction. Where non-competes are void, employers keep people with equity instead of contracts, so the deciding variable is your make-whole budget, not the law. Price the walk-away cost first, then decide whether you can cover it.

Golden handcuffs take several documented forms. Stock options or RSUs with four-year vesting at 25 percent per year. Retention bonuses payable after one to three years. Deferred compensation vesting at retirement or after five-plus years. Pension plans with five-year cliff vesting. Sign-on bonuses with one-to-two-year repayment clauses. Each has a different shape and a different exit cost.

The cost scales hard with seniority, and this is where a flat cash sign-on budget breaks.

The senior walk-away cost is about 10x the junior figure. Large acquiring employers handle this with formal buyout programs: Amazon, Google, and Meta match both the forfeited equity value and its vesting schedule, sometimes negotiating a discount. The make-whole structure practitioners describe is cash sign-on plus bridge RSUs, matched to the timeline, not a lump of cash against a multi-year grant.

Where non-competes are void, employers keep people with equity, so budget decides the hire, not the law.

Two tenure signals sharpen the financial score. First, position the candidate against the four-year vest and one-year cliff, because departures cluster post-vesting. The highest-yield contact window is the months just before a cliff, when the next tranche is close and the candidate has not yet reset. Second, check for a recent promotion. A promotion typically triggers refresh grants that reset the handcuff, so a long-tenured candidate who looks close to a cliff may have just restarted the clock.

This is exactly where a sourcing tool earns its place: finding the candidates who sit in a favorable financial window before you spend on any of them.

With Refolk you describe the friction profile you want in plain English - a state where non-competes are void, a tenure that puts the candidate near a vest cliff, a seniority band - and get back named people who fit, instead of assembling that filter by hand across public profiles.

The geographic dimension and the residence flip

Geographic friction is usually the smallest of the four, but it interacts with the contractual score in a way that can flip a verdict for free. Rate relocation need, remote allowance, and the residence-versus-HQ enforcement effect together.

The relocation part is straightforward: if the role requires a move the candidate will not make, that is a hard stop regardless of the other three scores. If the role is remote or the candidate is open to relocating, geographic friction is low.

The interaction is the interesting part. Because enforcement follows the work state, a candidate who lives and works in a void state carries any old restraint into a jurisdiction that ignores it. A senior engineer who relocated from Texas or Florida to California in the past two years is a clean example: whatever they signed at the old employer is almost certainly void where they now work. Reading residence against HQ can turn a contractual "no" into a "yes" at no cost.

The scoring procedure

Run the four dimensions in order, then combine. The full pass takes about half an hour per candidate and should happen before any message is drafted. There is a live disagreement about order: some practitioners score financial before contractual because unvested equity is the most common hard blocker in ban states, while others lead with contractual because it disqualifies fastest. Lead with whichever more often kills your candidates; the sequence below leads with the fast disqualifier.

The combine step is a judgement, not an average. Any single hard stop you cannot move - a required relocation the candidate refuses, a pre-I-140 visa stage with no near horizon, a senior equity load your budget cannot match - lands a deprioritize. Any single clearable barrier with a known date lands a pursue-later with that date attached. Only a clean sheet across all four lands a pursue-now.

Here is a verdict record you can paste into your CRM note field so the score survives being handed to someone else.

Candidate:
Employer / work state / HQ state:
Contractual: [void | threshold-dependent | enforceable] - basis:
Immigration: [portable | 180-day window | pre-I-140 | citizen/PR] - conditions checked:
Financial: est. unvested $______ ; refresh grant in last 12mo? [Y/N] ; make-whole feasible? [Y/N]
Geographic: [low | med | high] - relocation needed? [Y/N] ; residence flips contractual? [Y/N]
VERDICT: [pursue-now | pursue-later | deprioritize]
Next-touch date: ______ (tie to vest cliff or 180-day I-485 mark)

How this score goes wrong

The failure modes below are where a confident score is a wrong score. Each has a false positive and a cheap check. Read this section as the most important part of the standard, because an overclaimed friction score costs you good candidates just as surely as a missed barrier costs you an offer.

• Reading the choice-of-law clause instead of the work state. False positive: "candidate has a Texas non-compete, unmovable." Check: confirm where they physically work. A California-based worker's out-of-state non-compete is almost certainly void.
• Assuming a non-compete exists because the employer is known for them. False positive: scoring contractual friction high with no evidence. Check: prevalence is about 18 to 20 percent overall; only assume a restraint for executives, physicians, or trade-secret roles, where rates reach 45 to 67 percent.
• Treating the 180-day I-485 clock as the only immigration gate. False positive: "past 180 days, fully portable" while the I-140 is unapproved or your role is a different SOC. Check: both the I-140 approval and the same-or-similar occupation must also hold.
• Confusing the 180-day I-140 rule with the 180-day I-485 rule. Check: they are completely independent; verify each separately.
• Assuming any buyout makes the candidate whole. False positive: budgeting a flat cash sign-on against $500k of unvested RSUs on a multi-year schedule. Check: match both dollar value and vesting timeline, and account for the old stock's upside.
• Ignoring refresh grants after a recent promotion. False positive: scoring a long-tenured candidate as "near vest cliff, low friction" when a promotion just reset the handcuff. Check: look for a title change in the last 12 months.
• Assuming garden leave and non-solicits are void everywhere non-competes are. Check: California's treatment of garden leave is genuinely disputed and untested across reputable firms, and Washington still permits non-solicitation agreements targeting current customers.

The garden-leave point deserves weight because it is where writers overstate certainty. A garden-leave clause pays a departing employee full salary during a transition in exchange for not starting new work, and because they stay technically employed, some firms call it enforceable in California while acknowledging it has not been extensively litigated. Other firms state flatly that such policies will not be enforced. When the reputable sources contradict each other and no court has resolved it, treat garden leave as unresolved rather than scoreable, and flag it as a risk to check locally rather than a fixed barrier.

Before you call the score done

Verify the score with this checklist before you write the verdict. Skipping any line is how a barrier you missed at sourcing shows up at the offer stage.

Keeping the score current

The score depends on figures that move, so treat it as a living judgement, not a one-time verdict. Wage thresholds change every year: Washington's employee line moved from $120,559.99 to $123,394.17 to a cited $126,858.83 across three years, and Illinois holds at $75,000 until 2027 before rising. Statutory landscapes shift too - Washington's Engrossed Substitute HB 1155 is set to add it to the ban club on June 30, 2027. Re-check the state tier and threshold for any candidate whose score you carried over from a prior quarter.

The candidate-specific inputs decay faster than the law. A vest cliff arrives, a promotion fires a refresh grant, an I-485 crosses 180 days. Every one of those events changes the verdict, which is why pursue-later must always carry a dated trigger. When the reminder fires, re-run the two dimensions that moved rather than the whole score. The discipline that makes this framework pay off is not the initial scoring - it is returning to the calendared candidates on the exact date their friction drops, while everyone else is still guessing at the offer stage.