Building a portfolio case study that follows the entity, not the person

You will never get the accounts. That is the starting condition for anyone researching a privately held venture, and the useful response is not to guess at the numbers you cannot see but to become precise about the traces you can.

This is a method, not a set of conclusions. Portfolio case studies built this way work on any private business, and they work particularly well on ventures attached to well-known names, because those ventures generate a great deal of coverage and comparatively little evidence, which means the gap between what is asserted and what is documented is unusually wide and unusually easy to measure.

What to take away

The principle: follow the entity, not the person

A business is an entity. It has a legal name, a registration number, a jurisdiction, a status, and a paper trail. A person is a subject of coverage. Almost everything checkable attaches to the entity, and almost everything asserted attaches to the person.

So the first move in any research is to convert a claim about a person into a question about an entity: which company, registered where, and in what state. That split is exactly what matters for a famous name: the coverage keeps naming the person, while the register entry, the trademark filing and any secured lending record name a company that may not carry the person's name at all. If you cannot name the entity, you cannot verify anything, and that failure is itself a finding worth recording.

The evidence ladder

Rank every source you use. Not all of them are the same kind of thing, and conflating them is how confident errors get made.

The most common research failure is treating tier 6 as tier 1, quoting a modeled valuation as though it were a transaction. The second most common is circular sourcing: three articles that all trace back to the same press release are one source, not three.

Six public trails

1. The company register. Legal name, registration number, formation date, jurisdiction, status, registered agent, and depending on the country, directors, officers, shareholders, and filed accounts. In the United States that means the business registry of the secretary of state in the state of formation and the state of operation, and Delaware's Division of Corporations is the usual first stop for ventures organized as Delaware entities. In the United Kingdom it is Companies House. If a private company sits under a public parent, SEC EDGAR carries the parent's filings. The obligations that produce those entries are described in the guide to registering a business. Status is the most under-used field: administratively dissolved, struck off, forfeited, or delinquent tells you a company stopped filing, which usually means it stopped operating. In jurisdictions with public accounts filing, this trail alone can answer most questions.

2. Trademarks. The trademark database — the USPTO's Trademark Search system for searching, TSDR for a mark's full file — shows who owns the brand name as a mark, in which classes of goods, and with what status. In the United States, a mark filed on an intent-to-use basis becomes registered only after evidence of actual use in commerce is filed, so the presence of a registration, and the date of the use filing, is genuine evidence that something shipped. Assignments show transfers of ownership; recorded security interests show that the mark was pledged to a lender. Lapsed or canceled marks are a strong signal about a dormant brand, and the celebrity brands page explains why the license behind one usually ended first. Disputes over who owns a mark are decided at the Trademark Trial and Appeal Board, and its proceedings are public.

3. Secured lending records. In the United States these are UCC-1 financing statements and their amendments, filed with the secretary of state where the debtor is organized and searchable in that state's UCC index. They disclose that a lender has taken security over assets, who the lender is, and roughly what was pledged. Equipment financing, inventory financing, and factoring all leave this trace. The filing does not tell you the amount, but it tells you the business borrowed and against what.

4. Litigation dockets. Federal court records sit in PACER, the courts' public access system, and many state courts run their own public docket portals. Court records name the parties and contain the pleadings. Contract disputes, unpaid supplier claims, employment claims, and partnership disputes put private terms into public documents. Coverage of litigation is often thinner than the docket itself.

5. Regulatory and licensing records. Category-dependent and often the richest source. Food service permits are issued by the local health department under an adopted version of the FDA Food Code, and alongside them sit state alcohol control board licenses, inspections, health and safety records, product registrations in regulated categories, professional licenses, and import or export records. These name operators, addresses, and dates.

6. Operational footprint. Not documents, but observable facts, and the restaurant ventures page is the clearest case of how much they show: whether the product is actually stocked and where; how many locations exist and whether they are open; job postings and their seniority mix; the site's own history over time through a web archive; app store presence and update cadence; the range's expansion or contraction. Each is weak alone. Together they distinguish a functioning business from a landing page.

The date discipline

Every fact above has a date, and there are at least four different dates in play for any given claim: when the event happened, when the document was filed, when a source published, and when you looked. Conflating them produces most stale-information errors.

The two that matter most in this field:

Record all four dates next to any claim you rely on. It takes seconds and prevents the most embarrassing category of mistake.

What you cannot see, ever

The list below is the boundary of the method. These are not facts that are hard to find; they are facts no outside researcher can obtain from public records.

There is no research technique that reveals these for a private company. Estimates published elsewhere are models, and a model built on unavailable inputs is an opinion with arithmetic attached.

Building a case study that holds up

A defensible write-up has five parts, in this order.

The question. One sentence, narrow enough to answer. "Is this brand still trading?" is answerable. "Was this venture a success?" is not.

The entity. Legal name, number, jurisdiction, status, and date checked. If there are several related entities, say so and explain how they relate.

The record. What the documents show, each with its source and its date. Quote fields, not conclusions.

The gap. What you looked for and could not find, stated as plainly as what you did find. An absent trademark registration, a dissolved entity, a license with no renewal: absences are evidence when you have looked in the right place, and they should be described as such rather than left out.

The limits. What the evidence cannot support. If you have documented that a company exists and sells a product, say that. Do not extend it into a claim about who owns it, what it earns, or who profits.

Two rules of language make the write-up safe as well as accurate. Attribute claims to their source rather than asserting them ("the register lists the owner as..."), and describe entities rather than people when describing ownership. Both are more precise, not merely more cautious.

Putting the trails together in order

Take a celebrity-backed consumer brand that launched with substantial coverage and has since gone quiet. The shape varies — a spirits brand like Ryan Reynolds's Aviation Gin, a consumer goods line like Jessica Alba's The Honest Company — and so does how much becomes public: a venture that lists, as The Honest Company did, moves onto SEC EDGAR, while a privately held one stays on the state register. The outside-in sequence:

That sequence produces a defensible account of whether a business is trading, who holds its name, and whether anyone has lent against it. It produces no account at all of who made money, and it should not pretend otherwise.

Common questions

Is this worth the effort for an ordinary reader?

The first two steps, trademark and company register, take a few minutes each and are free. They answer more than most published coverage does.

What if the registers show nothing?

Then the venture may be foreign, may operate under a different legal name, may be a division of a larger company, or may not have existed as a distinct entity at all. Each is a real possibility and each is worth stating rather than resolving by assumption.

How do I handle conflicting sources?

Rank them on the ladder above and prefer the filed document. Where two documents conflict, prefer the more recent and note the conflict. Where a document conflicts with coverage, the document generally wins, but check whether they are describing the same entity.

Can I say who owns a company if the register lists an owner?

You can say what the register lists, on the date you checked. That is a different and more defensible statement than asserting ownership, because registers can be out of date, holdings can sit behind nominee and holding structures, and the registered owner of a trademark is not necessarily the beneficial owner of the business.

Related reading on this site: celebrity companies examples covers how to parse a specific ownership claim, and the celebrity companies portfolio guide explains the holding structures that sit between a person and an asset.