Refolk
FrameworkInvesting and deal sourcing

The Insider-Pass Read: Conviction Bridge, Capacity Gap, or Quiet Retreat

Score each existing investor on willingness and capacity from public and data-room evidence, then reach the round verdict another analyst would reach from the same file.

22 min readLast reviewed November 6, 2026Read as Markdown

Key takeaways

  • In Carta's fund data, every vintage from 2017 through 2020 had invested at least 89% of its available capital, so a pass from one of those vehicles says almost nothing about willingness.
  • The median 2022-vintage VC fund had deployed only 67% of its capital after nearly four years, against about 80% for other recent vintages, so a 2022 fund saying it has no money should be checked.
  • Score capacity against the vehicle that holds the shares, not the firm's newest fund, by matching the cap-table entity to the fund name in Form ADV Schedule D Section 7.B.(1).
  • Form D names no investors, and amendments for changes in amount sold or investor count are optional, so it cannot show whether insiders or new money filled a round.
  • In a 25-profile sample from Refolk's index of 800 US 'General Partner' profiles mentioning venture, only 11 still held that exact title, so checking that the sponsor is still there means confirming the current employer.
  • A board seat that is still filled does not prove the sponsoring partner stayed, because unless it was negotiated otherwise, the seat belongs to the firm and not the individual.

This guide is for investors offered an allocation in a startup's bridge or extension, and for leads pricing the round that follows one, who need to know why each existing backer sat out. You score every prior backer on two separate axes, willingness and capacity, using public records and the data room. Then you roll those scores into one round verdict: Conviction Bridge, Capacity Gap, or Quiet Retreat. The rules are tight enough that a second analyst working from the same file should reach the same verdict.

What does the insider-pass read decide?

It tells you whether existing investors' behaviour in a bridge says something about the company or only about their funds. A pass from a fund that cannot write the check looks the same on a cap table as a pass from a fund that has stopped believing. This framework is how you tell the two apart.

Most advice on reading insider participation in a bridge round treats it as one yes-or-no signal. That breaks down in exactly the cases that matter. An existing investor not taking its pro rata is a strong signal when it had money and chose not to spend it. It is close to noise when the vehicle ran out of money years ago.

Terms used throughout:

  • Bridge. Carta defines a bridge as any round raised after the first round in a series. Priced extensions count as bridges, not just SAFEs and notes.
  • Pro rata. An existing holder's right to buy enough of a new round to keep its ownership percentage.
  • Vehicle. The specific fund entity that holds the shares. Each vehicle has its own vintage, the year it started investing.
  • Sponsor. The partner who led the original check and usually holds or controls the board seat.
  • Willingness. Whether the backer wants to put more money into this company.
  • Capacity. Whether the vehicle holding the position is able to.

Every pass sits in one of four places

High willingnessLow willingness
Capacity Gap
Wants in but cannot fund. Discount the pass.
Conviction Bridge
Can fund and does. The insider signal is real.
Unreadable
Cannot fund and declines. Give the pass zero weight.
Quiet Retreat
Can fund and chose not to. The pass is the signal.
Low capacityHigh capacity
A pass only says something about conviction when the backer had the capacity to fund and chose not to.

The three round verdicts this guide produces:

  1. Conviction Bridge. Backers with money are putting it in. The people who know the company best are backing it.
  2. Capacity Gap. Most of the pro rata sits in vehicles that cannot pay. The insiders' behaviour tells you little about the company, so you are doing your diligence without an insider signal.
  3. Quiet Retreat. Backers who could have funded chose not to, or the prior lead had capacity and passed.

How do I score a backer's capacity?

Score capacity against the specific vehicle that holds the shares. Start from the typical remaining dry powder for that vehicle's vintage, then adjust for a confirmed reserve, a successor fund, and recent new leads. Dry powder is committed capital not yet invested. The score runs from C0 (unable) to C2 (able).

No fund administrator publishes the fund age at which firms stop leading new deals. The deployment curve is the only public proxy, and Carta's fund-admin data says how to read it. If a fund plans to reserve 20-30% of its capital for follow-ons, then by about 70% deployed it is basically done with net-new investments. What remains is reserves, spread across the whole portfolio.

The table below gives median undeployed capital by vintage from Carta's VC Fund Performance reports. The rows come from two different quarterly reports, so treat them as approximate and re-pull them from the latest report before relying on them.

VintageUndeployed (median)Carta reportStarting score
20177%Q3 2025C0
202011%Q2 2025C0
202117% (from 83% deployed)Q3 2025C1
202233%Q2 2025C2
202342%Q2 2025C2
202578%Q3 2025C2

Vintages missing from the table: Carta reports that every vintage from 2017 through 2020 had invested at least 89% of its available capital, so 2018 and 2019 start at C0. Carta puts year-one deployment for 2024 funds at 25%, so 2024 starts at C2.

Two readings anchor the table:

  • A pass from a 2017-2020 vintage carries almost no willingness signal. At 89% or more deployed, the fund is well past the point where net-new investing stops. Passing is the default whatever the fund thinks of the company.
  • A pass from a 2022 vintage is more informative than its age suggests. In Carta's Fund Economics Report, the median 2022 fund had deployed 67% of its capital after nearly four years, against about 80% for most other recent vintages. These funds hold more unspent capital than their age implies, so a 2022 fund saying it has no money deserves a closer look.

Then apply modifiers. Net them out, and keep the score between C0 and C2:

  • Up one if the backer confirms in writing that a reserve is set aside for this company, or names a successor fund that will write this follow-on.
  • Down one if the firm has no closed successor fund and no announced new lead in the last 6-12 months.
  • Down one if the sponsor's departure triggered a key-person clause that suspends new investment activity, until the backer confirms the suspension does not cover this follow-on.

Each public record proves something narrow. Here is what each one shows and how it can mislead:

  • Form ADV Schedule D Section 7.B.(1). Filed once per private fund, including by exempt reporting advisers, and it reports each fund's gross asset value. It proves the vehicle exists and how large it is. It does not show uncommitted reserves or how the fund allocates them, so a large figure can still hide an empty reserve.
  • The fund's own Form D. The amount-sold figure may include commitments, not cash. It can make a fund look more liquid than it is.
  • Close announcements. These prove commitments only. A fund can announce a close and still be calling capital slowly.
  • New-lead announcements. These show the firm is still writing first checks. They don't say which vehicle wrote them, so an active firm can still hold your company in a spent fund.

The slow part of this step is checking new-lead activity across a long backer list. In Refolk I ask in plain English for partners at seed funds with a 2023 or later vintage who led a new seed round in the past six months. Then I check the firm names that come back against my backer list. If a backer firm is missing from the results, I check it by hand rather than downgrading it automatically.

Is the sponsoring partner still at the firm?

Check whether the partner who led the original check is still at the firm and still holds the board seat, because the seat and the champion can come apart. A seat that is still filled does not prove the sponsor stayed.

An orphaned deal is a portfolio company whose sponsoring partner has left the firm. Kruze Consulting's practitioner view is that orphaned companies "typically have a much harder time getting follow-on capital inside the partnership." That is experience, not a measured effect, and no public dataset gives the size of the drop. It still matters here because it predicts passes for reasons that have nothing to do with the company.

What the sources say about the seat:

  • BetaKit: "Without further negotiation, the board (observer) seat belongs to the firm, not the individual."
  • Other practitioner sources note the designation right may belong to the fund, the firm, an affiliate, or in some cases the individual partner.
  • Some departing partners stay on as venture partners and keep the seat on the firm's behalf, so the same name on the board can hide a changed relationship.
  • Law-firm guidance, which I have only seen reported secondhand, warns that a seat may turn out to be personal to the departing partner.
  • A departure can trigger a key-person clause that suspends new investment activity.

How common is this? A head of platform quoted by CNBC said roughly one in three of their portfolio companies had at least one board member leave their firm. That covers one firm's portfolio, not the market. Still, it is frequent enough that you should check every backer.

Three public traces date a departure:

  1. The firm's team page, and whether the partner has dropped off it.
  2. A change in the partner's profile bio or current employer.
  3. The related-persons list in the portfolio company's latest Form D, which shows current directors but not when the change happened.

Record a sponsor status for each backer:

  • S1 Present. At the firm and holds or controls the seat.
  • S2 Departed, firm kept the seat. The seat is filled by someone else, or by the departed partner as a venture partner acting for the firm.
  • S3 Departed, seat personal. The designation right went with the partner.
  • S? Unconfirmed. Treat as departed until proven otherwise.

Do not run this check on titles alone. Title search over-counts current partners badly.

44%
of sampled "General Partner" matches still holding that exact title
In Refolk's index of professional profiles, 800 US profiles carry the title "General Partner" with the keyword "venture", and in a 25-profile sample only 11 had that exact current title, with others now listed as "Founder & CEO" or "Board Member".

That gap is why I confirm the current employer rather than the title. Refolk returns each sponsor's current employer next to their past roles. A partner who left a seed firm but still lists a board seat shows up as a departure, not a match.

How do I score willingness?

Only read willingness against capacity. A backer who could have funded and passed is the informative case. A backer who could not fund tells you almost nothing about conviction.

Score each backer once its position and capacity are known:

ScorePatternWhat it proves
W2Full pro rata or more, or leadsInformed money is backing it
W1Partial take-up, or out after confirmed cutbackPositive, limited by allocation
W0Out or partial with C1/C2, no cutbackChose not to spend money it had
WUOut with C0Nothing; cannot be read

Two rules keep this reproducible:

  • Ask the new lead whether it requested cutbacks. Investors also waive pro rata to make room for new investors. A pass made at the new lead's request is W1, not W0. Without confirmation from the lead, it stays W0.
  • Check the stated reason against the capacity evidence. If a 2022 vehicle says "we are fully deployed" and offers nothing to back it up, keep capacity at the table score. The pass then scores W0. A claim that the fund can't pay, made without proof, is not evidence.

The survey works best when the founder or lead runs it before the round is announced. Answers given after the announcement get shaped by who else has already committed.

Private pro-rata survey, one message per backer
Subject: Pro rata on the upcoming extension

Hi [name],

Before we open the extension to anyone else, I want your position privately.

1. Do you plan to take your pro rata: in full, partly, or not at all?
2. Which fund entity would write the check?
3. Is there a reserve set aside for us in that fund, and roughly how much is left?
4. If you are not taking it in full, what is the main reason?

One line on each is plenty. This stays between us.

[signature]

Send from the founder or the incoming lead before the bridge is announced; keep the four questions in this order.

A pass from a fund that cannot write the check tells you about the fund, not the company.

What can Form D tell me about who filled the round?

Form D can show cumulative sales and investor headcount on the date it was filed. It cannot show who bought, and amendments for those changes are optional. Treat it as a rough check on what you were told, never as proof of insider participation.

The issuer must file within 15 calendar days of the first sale. When an amendment is filed, it must give current information for every Form D field. The SEC also says an amendment is not required for changes such as the amount sold or the number of investors. So any change between filings is evidence you happen to get, not evidence the company has to give you.

FieldAmendment required on change?ProvesDoes not prove
Amount soldNoCumulative sales at filing dateWho bought
Investor countNoHeadcountNew versus existing holders
Offering amountYes, if up more than 10%UpsizingThat the round filled
Annual amendmentYes, if still continuingRound open after 12 monthsPace of sales
Related personsUpdated on filingCurrent directorsWhen the change happened

Filing dates also drift. Of 29,947 Form D notices that give a first-sale date, 59.8% were filed within the 15 days the rule allows and 40.2% were filed later. Always date the round from the first-sale date, not the filing date.

Evidence on who filled the round, strongest first

  1. Data-room closing schedule
    Names, amounts, and dates for each investor in the round
  2. Backer's stated position
    In, partial, or out, with a reason, given privately
  3. Public capacity records
    Vintage, ADV fund entry, successor fund, recent new leads
  4. Form D deltas
    Cumulative sales and headcount, no names, amendments optional
Only the data room names who bought; every layer below it narrows the question without answering it.

The insider-pass procedure, step by step

Run these seven steps in order. The full read takes about two weeks of elapsed time, mostly waiting on the survey, and roughly half a day of analyst time per backer.

From cap table to round verdict

  1. Build the backer list by vehicle
    List every holder with pro-rata rights and the fund entity that holds the shares, with its vintage. Done means one row per backer with the vehicle named, in about a day.
  2. Survey pro-rata intent privately
    Have the founder or lead ask each backer for its position before the round is announced. Done means every backer has stated in, partial, or out with a reason, usually in one to two weeks.
  3. Score capacity from public records
    Start each vehicle at its vintage score, then adjust for a confirmed reserve, a successor fund, and new leads in the last 6-12 months. Done means C0, C1, or C2 per backer, at about half a day each.
  4. Check the sponsor
    Confirm the leading partner is still at the firm and still holds the seat, using team pages, bio changes, and the company's latest Form D related persons. Done means a sponsor status per backer, at one to two hours each.
  5. Score willingness against capacity
    Compare each stated position with the capacity score and ask the new lead about cutbacks. Done means separate W and C scores per backer, with passes from C0 vehicles marked WU.
  6. Reconcile with Form D
    Compare amount sold and investor count on any filing against the commitments you were told about. Treat it as weak evidence and finish in under an hour.
  7. Roll up to the round verdict
    Weight backers by pro-rata dollars and apply the three verdict rules in fixed order. Done means one verdict with its two ratios written beside it.

On step 2, sources disagree about order. One practitioner source says to go to existing investors almost always first. Another treats existing investors first versus new lead first as an open sequencing decision. The framework doesn't depend on the order. It only needs every position collected before the round goes public.

How do backer scores roll up into a round verdict?

Weight each backer by its pro-rata dollars in this round, then apply three rules in a fixed order. The first rule that fires is the verdict. Fixing the order is what lets a second analyst reproduce the result.

Compute four quantities from the scoring sheet:

  • E: total pro-rata entitlement in this round, in dollars, across all backers.
  • E-able: entitlement held by backers scored C1 or C2.
  • E-unable: entitlement held by backers scored C0.
  • P-unexplained: dollars of E-able not taken up, excluding amounts covered by a confirmed cutback request.

Apply in order:

  1. Quiet Retreat if the prior round's lead scores C1 or C2 and W0, or if P-unexplained is 50% or more of E-able.
  2. Capacity Gap if E-unable is more than 50% of E.
  3. Conviction Bridge otherwise.

If E-able is zero, rule 1 cannot fire and the verdict is Capacity Gap. The 50% thresholds are a convention I use for consistency, not an empirically derived cutoff. Keep them fixed across deals and write both ratios next to the verdict, so a reader can see how close the call was.

VerdictFires whenWhat the passes tell meBefore pricing
Quiet RetreatAble lead passed, or half of able pro rata unexplainedInformed money chose not toGet each reason in writing; reprice or pass
Capacity GapOver half of pro rata in C0 vehiclesLittle about the companyDiligence as if no insiders exist
Conviction BridgeNeither rule firedInformed money is backing itDiligence anyway; bridging is still a selection signal

A Capacity Gap verdict has a second consequence that people tend to skip. You will inherit a cap table where most prior backers cannot support the next round either. Plan that round's syndicate without them.

Insider-pass scoring sheet, one row per backer
Backer | Holding vehicle | Vintage | ADV 7.B.(1) match (Y/N) | Successor fund closed (Y/N) | New lead last 6-12 mo (Y/N) | Reserve confirmed (Y/N) | Capacity (C0/C1/C2) | Sponsor status (S1/S2/S3/S?) | Key-person triggered (Y/N/unconfirmed) | Pro-rata entitlement ($) | Taken up ($) | Cutback confirmed ($) | Stated reason | Willingness (W2/W1/W0/WU)

Roll-up:
E = sum of entitlement
E-able = entitlement where Capacity is C1 or C2
E-unable = entitlement where Capacity is C0
P-unexplained = (E-able minus taken up by C1/C2 backers) minus cutback confirmed
Rule 1 Quiet Retreat: prior lead is C1/C2 and W0, or P-unexplained / E-able >= 50%
Rule 2 Capacity Gap: E-unable / E > 50%
Rule 3 Conviction Bridge: otherwise
Verdict: ______  P-unexplained / E-able: ___%  E-unable / E: ___%

Paste into a spreadsheet; fill every column before computing the roll-up, and leave nothing blank (use "unconfirmed").

How does the insider-pass read go wrong?

Most wrong verdicts come from a handful of repeatable errors: reading the wrong unit, the wrong date, the wrong entity, or the wrong person. Each one below names the false positive and the check that catches it.

Treating every bridge as distress

Carta counts priced extensions as bridges, so "this is a bridge" is not the same as "this is a rescue." Check the instrument actually used: a SAFE or note, or priced preferred stock. A priced extension with a lead at a step-up is a different event from a note issued to buy time.

Reading a headline bridge share as a share of deals

Carta's headline bridge figure can measure dollars, not deals. Carta data showed bridges at about 16.6% of capital raised on its platform in one quarter, up from 11.8%. That is a share of dollars. Before you quote any bridge prevalence figure, check whether it counts cash or rounds.

Treating a missing Form D amendment as a stalled round

Amendments for amount sold are optional. A stale filing is the normal case. Get the closing schedule.

Treating a late Form D as a recent round

With 40.2% of notices filed after the 15-day window, the filing date can lag the round by months. Use the date of first sale.

Taking the vintage from the firm instead of the vehicle

A firm that just closed a new fund may hold your company in an older, fully deployed vehicle. The firm looks able, but the vehicle is C0. Match the holding entity on the cap table to the ADV 7.B.(1) fund name.

Assuming the champion stayed because the seat did

The seat may sit with the firm after the sponsor has left. A departing partner may also keep it as a venture partner acting for the firm, so the same name on the board can hide an orphaned position. Check the team page and the related persons in the latest Form D.

Inferring willingness from capacity alone

A backer with capacity that passes is not automatically W0. Investors waive pro rata to make room for new investors, sometimes at the new lead's request. Ask the lead whether it requested cutbacks before you score the pass.

Counting active partners from a title search

In the Refolk sample above, 11 of 25 title matches still held the exact "General Partner" title. A sponsor check built on titles alone will mark departed partners as present. Verify the current employer.

Reading the bridge graduation gap as cause

Peter Walker, Head of Insights at Carta, reported that only 4% of 2020 seed companies that raised a SAFE or note bridge after their seed have reached Series A, against about 50% of those that did not bridge. Secondary write-ups of a Carta chart give 7% to 14% graduation for bridged companies against 21% to 29% for those that did not bridge. I did not locate the primary chart.

4%
of 2020 seed companies with a SAFE or note bridge that reached Series A
Against about 50% for 2020 seed companies that did not bridge, per Carta's Peter Walker; the gap mostly reflects which companies needed a bridge, not what the bridge did.

Companies that bridge on SAFEs had usually already missed the Series A bar, so the gap is mostly selection. The baseline also moves a lot by cohort. Every figure below is a secondary transcription of Carta data:

Seed cohortReached Series A within 2 years
Q1 201830.6%
201927.5%
H1 2020~40%
2022~17%
Q1 202315%

No public data splits graduation by insider-led, mixed, or outside-led bridges, or by first versus second bridge. If someone tells you that insider-led bridges graduate at a known rate, that claim has no public source. The verdict here is a read on insider behaviour, not a forecast of whether the company reaches Series A.

Before you price: final checks and keeping the read current

Run the checklist before the verdict goes into an investment memo. Then refresh the inputs that change over time: deployment curves move every quarter, and partners move without notice.

Insider-pass read, ready to price

  • Every backer with pro-rata rights has one row, with the holding vehicle named rather than the firm.
  • Each vehicle's name matches a fund entry in ADV Schedule D Section 7.B.(1), or the mismatch is noted.
  • Each capacity score starts from the vintage table and lists every modifier applied.
  • Every backer has a stated position (in, partial, or out) gathered before the round was announced.
  • The new lead has confirmed in writing whether it requested any cutbacks.
  • Every sponsor status was set from the current employer, not a title match, and S? entries are treated as departed.
  • Every pass from a C0 vehicle is marked WU and none is counted as a conviction signal.
  • Round dates use the Form D first-sale date, and the closing schedule came from the data room.
  • The verdict sheet shows P-unexplained / E-able and E-unable / E beside the verdict.

To keep the read current:

  • Deployment curves. Re-pull the vintage table from the latest Carta VC Fund Performance or State of Private Markets report each quarter. Re-score any backer whose vintage figure has moved enough to change its starting score.
  • Sponsor status. Re-check team pages and current employers within a week of signing. A departure between survey and close changes both capacity, through key-person clauses, and how you read the pass.
  • Open offerings. If the round is still open a year after first sale, the annual Form D amendment is mandatory. That filing is your next free look at cumulative sales and headcount.
  • Base-rate claims. Before you cite any figure on insider-led bridge share or insider discounts, look for the primary source. The commonly quoted versions are unverified.

Questions practitioners ask

Is it a bad sign if the lead investor doesn't take its pro rata in a bridge?

It depends on whether the lead's fund could have paid. If the holding vehicle is from a 2017 to 2020 vintage, which Carta shows at least 89% invested, a pass is the default and tells you little. If the vehicle is from 2022 or later and the lead passed without the new investor asking for cutbacks, that is the most informative pass on the cap table. Under this framework it triggers a Quiet Retreat verdict on its own.

How do I tell if the lead investor's fund is out of reserves?

You cannot see it directly in public records. Form ADV Schedule D Section 7.B.(1) shows that a fund exists and its gross asset value, not uncommitted reserves. Use the vehicle's vintage against Carta's deployment curves as the proxy, check for a closed successor fund and recent new leads, and then ask the backer to confirm in writing whether a reserve is set aside for this company.

What happens to follow-on support when the partner who led the deal leaves?

It usually gets harder. Kruze Consulting's practitioner view is that orphaned companies typically have a much harder time getting follow-on capital inside the partnership, though no public data measures the effect. Unless negotiated otherwise, the board seat stays with the firm, and the departure can trigger a key-person clause that suspends new investment activity. Confirm the sponsor's status and whether any suspension covers this follow-on.

Can Form D show whether existing investors participated in a round?

No. Form D reports cumulative amount sold and investor count, never investor names, so it cannot separate insiders from new money. Amendments for changes in amount sold or investor count are also optional, so a missing amendment does not mean the round stalled. Use Form D only as a rough check against the commitments you were told about, and get the named closing schedule from the data room.

Do insider-led bridges reach Series A more often than outside-led ones?

That is not established publicly. Carta's Peter Walker reported that only 4% of 2020 seed companies that raised a SAFE or note bridge reached Series A, against about 50% of those that did not bridge. No public dataset splits that gap by insider-led versus outside-led bridges, and the gap mostly reflects which companies needed a bridge in the first place, not what the bridge itself did.

Should the founder go to existing investors or a new lead first?

Sources disagree. One practitioner source says almost always existing investors first. Another treats the order as an open sequencing decision that depends on the round. For this framework, the order matters less than the timing: get every backer's private pro-rata position before the bridge is announced, so that what they say isn't shaped by who else has already committed.

Try it on the search you came here for

Stop building boolean strings. Just describe the person.

Type one sentence. I plan the search, read GitHub, public LinkedIn and Crunchbase records, and the open web as it is right now, and hand back a ranked list with the reason next to every name.

  1. 01Describe them

    One plain sentence. Role, city, stack, stage, whatever matters to you.

  2. 02I read the web live

    GitHub, public LinkedIn and Crunchbase records, the open web. Not a database that went stale last quarter.

  3. 03You read the shortlist

    Ranked, with the reasoning under every name. Open a profile, ask a follow-up, narrow it down.

  • No boolean, no filters, no seat to buy. One box.
  • Read at search time, so a profile updated yesterday counts today.
  • Every step visible as it runs, every name with its reason.

500 free credits on sign-up. No card, no demo call. See real searches.

Read next