The Re-Founder Readiness Score: Timing an Exited Founder's Next Company

This guide is for early-stage investors, platform and talent partners, and angels who want to reach founders who sold their last company before those founders raise their next round. The job is to decide which exited founders are close to starting again and rank them for outreach. What this delivers is a repeatable score: 0 to 100 across five weighted dimensions, an estimated free-to-build date attached to each name, and three bands - reach now, watch, not yet - you can apply to a case in front of you.

Existing sourcing playbooks read a pre-launch footprint, chase spinout founders, or map a warm intro. None of them estimate the retention clock that governs when a sold founder is actually free to build again. That clock is the spine of this score. Get it right and you reach people in the window between vest-out and first filing. Get it wrong and you either pitch someone who is two years locked or find them the same week their launch hits the press, alongside every other fund.

Why a readiness score, and not just a stealth-bio search

A readiness score beats a stealth-bio search because a bio is a lagging, noisy signal and the retention clock is a leading, computable one. Most guides tell you to watch for a headline that flips to "Building something new." By then you are late, and half the people who write that never ship a company.

The reason to score instead of scan is timing. By the time a stealth startup surfaces through a launch or a funding event, the founder has already chosen a lead investor and locked first hires. The teams that flagged it reached the founder weeks or months earlier. Founders build investor relationships ideally 6 to 12 months before they plan to raise, so the outreach that matters happens well before any public event.

The between-companies pool is also larger and more legible than the stealth pool. In Refolk's index, about 1,978 US profiles currently hold Entrepreneur in Residence or Founder in Residence titles, a common holding pattern for exited operators between companies. On a deliberately narrow query combining title and keyword, only 7 US profiles use explicit stealth or "building something new" language, against 1 in the UK.

That contrast is the whole argument. Waiting for someone to declare stealth throws away most of the addressable market. Scoring the retention clock across an acquirer list surfaces people who have not declared anything yet.

The five dimensions and how to weight them

The score has five dimensions. No published standard scores re-founder readiness this way, so the weights below are mine, anchored on documented norms, not on a validated model. Treat them as a defensible starting rubric to calibrate against your own hit rate.

The retention clock carries the most weight because it is the one dimension the other guides ignore and the one grounded in the hardest data. Disengagement and build signals split the next largest share evenly, because either alone lies and both together corroborate. Role fit and serial intent are tie-breakers that separate two founders with identical clocks.

Score each dimension 0 to 100, multiply by its weight, and sum. A founder who is a year from vest-out with no build signals scores low even if their bio says stealth. A founder three months past a completed earn-out with a founding-engineer posting and a fresh GitHub org scores high even if their headline still names the acquirer.

The retention clock: reading the earn-out math

The retention clock is the estimated date a founder is contractually and financially free to build again, computed as close date plus the applicable earn-out or re-vesting window. It carries the most weight because it is computable from primary records and because departure tracks it tightly.

Documented transition and retention periods run 1 to 3 years, with strategic tech acquirers pushing 2 to 4 years to release full payout. PE buyers are often more flexible because they have replacement CEOs ready. Around 40 percent of total founder consideration is commonly held back through escrow, retention grants, and reverse vesting for 2 to 3 years. Reverse vesting alone commonly re-vests 20 to 40 percent of already-vested shares.

Here is the earn-out distribution to anchor your default.

Source: aggregated M&A post-closing data. A separate baseline of 735 private deals from 2012 to 2015 found 47 percent had earn-outs of one to two years, while 21 percent ran four to five years.

The reason the clock dominates is that departure is a step function, not a slope. Attrition jumps from 12 percent of comparable regular hires to 34 percent of acquired workers in year one, and reaches roughly 90 percent gone once the earn-out ends. That is why time-since-close should be scored as a threshold crossing at the earn-out expiry, not as a linear countdown.

Score the retention-clock dimension high when the founder is at or past the estimated free date, moderate within three months of it, and low when they are more than six months out. The three-month leading band matters because filings lead announcements: Form D, incorporation, and trademark records typically predate public mention by a full quarter.

Fixing the close date from primary records

The close date is the anchor of the whole clock, and the strongest primary record for a US public acquirer is the Form 8-K. The acquirer reports entry into a material acquisition agreement under Item 1.01 and discloses the closing under Item 2.01. The closing 8-K is time-bound: it must be filed no later than four business days following the Closing Date.

That makes the Item 2.01 date high-reliability, but only for deals above the significance threshold and only for SEC filers. Two cautions follow directly.

• Item 1.01 is the signing, not the closing. Using the signing date under-counts the clock and pulls a founder into "reach now" early. Verify which item.
• Small or foreign deals never trigger an 8-K. Absence of a filing is not absence of a deal. Fall back to press release, incorporation and registry records, or Form D filings, and lower your confidence.

For private acquirers, hold two dates: the approximate close from press and a wider confidence band on the free date. A founder whose free date could land anywhere across a twelve-month spread belongs in watch, not reach now, regardless of what their bio says.

Reading disengagement and build signals

Disengagement and build signals corroborate the clock: the clock tells you when a founder can leave, these tell you whether they are actually moving. Neither is trustworthy alone, which is why they split the middle weight and why the rule is two concurrent signals before flagging.

Disengagement is about pulling away from the acquirer. The reliable public tells:

• A LinkedIn headline shifting from a senior role to "Building something new" or "Stealth," dated within the last few months.
• A tenured operator leaving without announcing a next step, especially if two teammates leave within weeks.
• Profile edits and post activity that no longer reference the acquirer, even while the current-employer field still does.

That last point is the trap. A founder still listing the acquirer may already be checked out - the rest-and-vest state. Cross-check post recency and title edits, not just the current-employer field, or you will misread a disengaged founder as committed.

Build signals are about starting the next thing. The precursors, in rough order of strength:

• A founding-engineer job posting at a company with no website. This is one of the strongest signals because it commits money and intent.
• A Form D, fresh incorporation record, or trademark application, which often predates public mention by a full quarter.
• A registered domain or a bare landing page.
• A sparse new GitHub organisation.

The signal that actually converts is a hire, not a bio, because a hire spends money. Weight a founding-engineer posting at a website-less company above any number of profile edits.

Assembling that list by hand means cross-referencing acquisition records, exit dates, job postings, and current employers across several sources. Refolk runs the whole pattern from one plain-English prompt: acquirer and exit constraints, time-since-close, and a live build signal in a single query, so you skip the manual join and go straight to the shortlist.

The scoring procedure, end to end

Run the seven steps below in order. Each step feeds the next, and the whole thing takes about a day for the list-build and minutes per founder thereafter.

Steps one through three are the acquirer-list-in, ranked-builders-out pattern. One demo mined 806 acquisitions across six acquirers and ranked 736 people, which is the shape of the top of funnel you are building. Live-verification at the end is not optional: in one vendor demo re-run, 19 acquisitions in produced 13 live-verified founders out, so expect to lose roughly a third to stale or mismatched profiles.

Turning the score into reach now, watch, or not yet

The band is the decision the score exists to produce, and it is a two-variable judgement: how close the free date is, crossed with how strong the concurrent build signals are. The thresholds below are mine and untested, so treat them as a sort order to calibrate, not a probability.

As a rough numeric mapping: score 70 and above with a free date reached or within three months goes to reach now. Score 40 to 69, or a strong signal with a distant clock, goes to watch. Below 40, or a founder more than six months from free with no signals, is not yet.

The re-founding metronome is real and worth calibrating against. Sqreen sold for roughly $260M in 2021, and both founders were building Tolmo by 2025, a four-year interval that mirrors a standard four-year vesting package. A demo on the six most recent Datadog deals found them still fully staffed, exactly as four-year vesting predicts. Standard founder vesting is four years with a one-year cliff, and nearly all VCs require it, so a founder acquired inside the last two years is almost always in the not-yet band no matter how their bio reads.

The clock tells you when a founder can leave; the build signals tell you whether they already have.

How this score goes wrong

The score fails in seven documented ways, and every one of them produces a false positive that wastes an outreach slot or a false negative that hands the deal to a faster fund. Read this section as carefully as the rubric.

• Stale free-date default. Assuming a flat two-year clock misses the 4-to-5-year earn-outs, which were 21 percent of one dataset. You score someone "reach now" who is still two years locked. Check the actual 8-K or press terms before trusting the default.
• Stealth headline that never converts. A "Building something new" bio can be a sabbatical or an advisory pause, or an EIR parked at a fund. Require a second concurrent signal - domain, Form D, or founding-engineer post - before flagging.
• Rest-and-vest misread as engaged. A founder still listing the acquirer may already be checked out. Cross-check post activity and title recency, not just the current-employer field.
• 8-K date confusion. Item 1.01 is the signing; Item 2.01 is the closing. Using the signing date under-counts the clock and pulls people into reach now too early. Verify which item.
• Sub-threshold and private deals invisible. Small or foreign deals never trigger an 8-K, so absence of a filing is not absence of a deal. Fall back to registry and press, and lower your confidence.
• Single-signal false alarm. One domain registration is noise. Require two or three signals in the same window on the same person before flagging.
• Score theatre. The thresholds are author-defined with no published false-positive rate, so a precise number implies false confidence. State the assumption and back-test on known cases.

The common thread is that a single strong-looking signal, read in isolation, is the most reliable way to mis-time a founder. The clock corrects the bio, the build signals correct the clock, and live-verification corrects all three.

Verification checklist before outreach

Run this before any name leaves the list. It is the difference between a queue you can trust and a queue that reaches locked founders.

Keeping the queue current

A readiness queue decays because clocks tick and signals appear, so re-run it on a cadence rather than treating it as a one-time build. The highest-value window sits between vest-out and first filing, and that window opens and closes on its own schedule for every founder on the list.

Set two triggers. First, a monthly recheck of every watch-band founder whose free date falls inside the next quarter, looking for the first build signal that promotes them to reach now. Second, a standing search on your target acquirers for founders who have just crossed their estimated free date, using a query such as founders who sold a company to a named acquirer in a given window and no longer list that acquirer as their current employer. A tool like Refolk lets you save that acquirer-and-timing query and re-run it, so the top of funnel refreshes itself instead of decaying between manual pulls.

Back-test the thresholds as known cases resolve. When a founder on your list raises publicly, record where they sat in your bands the quarter before the round. Over a few cycles that gives you the one thing the published literature does not: a false-positive rate for your own cut-offs, which is what turns this from a defensible rubric into a calibrated one.