Verifying an Emerging Manager's Deal Attribution From Public Records

You are an LP or an angel looking at a first-time fund manager's track-record slide, and the slide says "I led these deals." Before you spend a reference call testing that claim, you can settle most of it from public records. This guide carries one track-record slide all the way through the public-records pass and shows you how to classify each logo as verified-led, verified-participated, timeline-corroborated-only, or unsubstantiated, so you arrive at your back-channel calls with the claims already sorted and your call time reserved for what public evidence cannot decide.

The library already has a back-channel reference playbook and a target-LP-list guide for the raising side. This is the other half: the public-records pass an allocator runs before picking up the phone.

Why the public-records pass comes first

Run the public trail before any reference call, because public filings settle the cheap questions and leave only the expensive ones for humans. Nearly three quarters of institutional LPs already review the data room before their first GP meeting, so the ordering is not radical. What is new here is treating the public-records pass as its own verdict-producing stage rather than a footnote.

The two ordering conventions disagree. SVB's emerging-manager guidance frames public data as a fallback you reach for when attribution letters are missing. The LP-diligence approach I use here puts the public-records pass first, before a single call. I prefer public-first for one reason: reference calls are your scarcest resource, and a call spent confirming that a round happened during the GP's tenure is a call wasted. Filings answer that. Save the human for the questions only a human can answer.

For a first-time manager, LPs focus on individual deal attribution, loss ratios, and relevance to the stated strategy. Attribution is where the slide does the most work and carries the most risk, so that is where the public pass earns its keep.

The five attribution tiers and which ones filings can prove

The standard track-record slide claims one of five roles per deal, and only one of them leaves a reliable public footprint. SVB and the derivative LP guidance define the tiers as sourced, co-led, led, board member, and staffed. The recommended slide format lists company name, sector, entry and exit dates, the GP's attribution among those five verbs, co-investor names, and gross and net multiples deal by deal.

Here is the structural fact that shapes everything downstream: of those five verbs, only "board member" can be evidenced from public records, and only sometimes. The other four require attribution letters or references.

The ceiling is structural, not a matter of effort. Form D names only the issuer's officers, directors, and promoters, and never the investors, so "led," "co-led," and "sourced" are unprovable publicly by construction. More searching cannot upgrade those tiers. This is the single most important thing to internalise before you start: you are not trying to prove "led" from filings. You are trying to disprove it, or bound it, or push it down to a tier the evidence supports.

What each public source can and cannot settle

Three public sources carry the load: Form D on EDGAR, Schedule 13D/13G, and IAPD. Each proves a different thing, and confusing them is the most common way this work goes wrong. Fix their scope before you touch a single logo.

Form D discloses the issuer's name and state, related persons (officers, directors, promoters), industry, size range, exemption claimed, date of first sale, total offering amount and amount sold, and number of investors. It must be filed within 15 days after the first sale, and once filed it generally cannot be withdrawn, with no confidential treatment available. That permanence is useful: the timeline it fixes is hard to revise later. It omits investor identities, price per share, and any board designee who is not also a listed officer or director.

Schedule 13D applies only to public companies. It is triggered at 5% or greater beneficial ownership of a class of equity, and it can state board membership directly, showing a named person becoming a member of the issuer's board. A passive holder who files the short-form 13G must switch to 13D upon seeking a board seat, so a 13D is a strong positive signal. But its absence proves nothing about a private-round seat.

IAPD (Investment Adviser Public Disclosure) confirms adviser registration and history: current registrations, employment history, and disciplinary events. It fixes the GP's identity and tenure. It says nothing about any specific deal. IAPD retains records for ten years after an adviser is no longer registered, so a departed GP's prior-firm dates remain checkable.

The soft-title reservoir that collapses under this pass

The titles that read as "led" on a slide often carry no filing footprint at all, and the volume of those titles is large. In Refolk's index there are 4,717 US profiles carrying a "Venture Partner" title and 3,648 carrying a "Board Observer" title. These roles lend credibility to a track record and leave nothing to verify.

A venture partner is frequently a part-time or deal-by-deal affiliation, not a full-economics GP role. A board observer attends and listens but does not vote, and observer seats rarely appear in any filing. Both collapse to timeline-corroborated-only under a public-records pass, because there is no artifact to lift them higher. When a slide says "led" but the GP's public profile reads "Venture Partner" or "Board Observer" at that firm, you have found your first fork.

Resolving those profiles by hand across LinkedIn, firm pages, and filings is slow. Asking Refolk in plain English for the people who carry a given title at a given firm surfaces the soft-attribution layer in one pass, which is where the "led" claims that will not survive tend to hide.

The worked example: one slide, four logos

Take a first-time manager's slide with four logos and the verb "Led" under each. Here is the public trail I ran and the fork I hit at each one, including the deal that looked led and turned out to be a fourth-name observer seat.

Logo 1, a company that later IPO'd. IAPD fixed the GP's CRD number and confirmed they were at the prior firm during the claimed vintage. Form D corroborated the round: right first-sale date, right size, GP not listed as a related person. Then the 13D. The company went public, a holder crossed 5%, and the 13D named the GP as becoming a member of the issuer's board. That is a documented voting seat. Verdict: verified board member. The "led" verb still is not proven, but the seat is, and a board seat is a strong floor.

Logo 2, a still-private company. IAPD confirmed tenure. Form D corroborated the round timing and size, and the number of investors was disclosed as a small figure consistent with a lead-plus-a-few round. The GP was not a Form D related person. No 13D exists, because the company is private. The press release quoted the GP. That is presence, not role. Verdict: timeline-corroborated-only. The vintage overlaps, the round is real, the GP was quoted, but nothing documents that the GP set terms or took a seat. This one goes to a reference call.

Logo 3, the one that looked led. The slide said "Led." IAPD confirmed tenure. Form D corroborated the round. No 13D, private company. But the firm's own portfolio page listed the GP under "Board Observer," and a co-investor's press release described a six-person deal team on which the GP was the fourth name. This is the failure mode in the flesh: someone who showed up to two meetings as the fourth person on a six-person team. Verdict: unsubstantiated as led, correctly a board observer at most. This is the logo a reference call would have wasted an hour confirming; the public trail settled it.

Logo 4, a company the GP joined as an officer. Form D listed the GP as a related-person director. Tempting to read as a board seat. But the address on the Form D tied to the issuer, not the fund, and IAPD showed the GP had an operating title at that company before the fund existed. The GP was a company officer, not the fund's board designee. Verdict: not an investment attribution at all. It belongs on an operating resume, not a track-record slide.

Run it yourself: the procedure

Work one GP's slide top to bottom. The public-records pass runs about a day for a slide of eight to twelve logos, front-loaded on the Form D pull.

Use this scoring template to record each logo so the verdicts are auditable when your IC asks how you got there.

Company | Claimed tier | GP on Form D? | 13D board seat? | Press role | CRD/tenure match | Verdict | Evidence
Logo 1  | Led          | No            | Yes (director)  | Quoted      | Yes              | Verified board member | 13D accession no.
Logo 2  | Led          | No            | N/A (private)   | Quoted      | Yes              | Timeline-corroborated-only | Form D + release
Logo 3  | Led          | No            | N/A (private)   | Observer    | Yes              | Unsubstantiated as led | Portfolio page
Logo 4  | Led          | Yes (officer) | N/A             | None        | Operating title  | Not an investment claim | Form D address

The verdict language matters. "Verified-led" is a tier you will almost never assign from public records, and that is correct. "Timeline-corroborated-only" is the honest resting place for most private-company logos. It tells your IC exactly what the public trail settled and exactly what a reference call still needs to establish.

How this goes wrong: the seven failure modes

The public-records pass produces confident-looking verdicts, which is precisely why its failure modes are dangerous. Each one below is a specific way a filing lies or is misread. Learn the tell for each.

• Form D "director" false positive. A GP named as a related person may be a company officer, not the fund's board designee. The tell: check whether the person's address ties to the fund or the issuer. Logo 4 above was this exact trap.
• Board observer read as board member. The deal that looks led is often a fourth-name observer seat. Observer seats rarely appear in any filing, so the tell is the absence of a 13D or proxy director listing combined with a portfolio-page "Observer" title. Do not read silence as a voting seat.
• 13D mismatch (false negative). A 13D exists only once a company is public and a holder crosses 5%. Using its absence to disprove a private-round board seat is a false negative. For private companies, 13D can only ever add evidence, never subtract it.
• Common-name collision. Two advisers share a name. Resolve only via CRD number plus overlapping employment dates, never name alone. Collision risk rises in small markets: with 9 UK versus 146 US GP-titled venture profiles in Refolk's index, a UK match has far fewer disambiguating peers, so CRD-plus-dates resolution is more load-bearing there.
• Press-release inflation. A quote in a financing release proves presence, not that the GP led. Treat every quoted-in-release logo as timeline-corroborated-only until a role is documented elsewhere.
• The timeline-only trap. A Form D confirms the round happened during the GP's tenure. It does not confirm the GP touched that deal. Do not upgrade to verified-participated on vintage overlap alone.
• Reclassified categories. Watch for a tier or sector shifting between pitch versions. This is the documented pattern behind an April 2020 SEC case that fined a manager one million dollars for reclassifying a fund investment to inflate a track-record category.

There is also a fraud layer that no clean public trail rules out. The SEC's September 2024 Crawford Ventures order documented a fabricated "Performance Audit" attributed to a real audit firm that never did the work. Public filings can corroborate a round; they cannot vouch for a slide's honesty. Where the manager is an exempt reporting adviser, this matters more, because ERAs sit outside the Marketing Rule and face only anti-fraud enforcement. The guardrail on performance claims is thinnest exactly at Fund I, which is why the burden falls on you.

Before you call it done

Verify these before you promote any logo above its evidence or hand the slide to your investment committee. This is the gate between a public-records pass and a reference call.

Keeping the read current and honest

The public-records ceiling does not move, but the trail around a manager does. A still-private logo can go public and produce a 13D that upgrades a timeline-corroborated-only verdict to a verified board seat. A manager can outgrow ERA status and become a registered adviser fully subject to the Marketing Rule, which changes how much you can lean on the slide's own numbers. Re-pull the Form D and IAPD trail before each closing, not just at first look.

The public-records pass is not a substitute for reference calls. It is the thing that makes reference calls worth their cost. When you walk into a call already knowing that Logo 1 is a documented board seat, Logo 2 is a real round with an unproven role, Logo 3 is an observer seat dressed as a lead, and Logo 4 does not belong on the slide, you spend the entire call on the one question that matters: what did this person actually do on the deals the filings could not decide. That is the whole point of running public records first.

Keep the per-logo worksheet with your IC memo. When a co-investor or a portfolio CEO surfaces later as a reference, the worksheet tells you exactly which claim they can and cannot corroborate, and you avoid asking a founder to vouch for a role the public record already settled.