Researching Obscure Figures in Public Records
I spent about three weeks last year digging into a name that barely appeared anywhere online. The search started because someone referenced a brand and a woman connected to it, but the digital footprint was fragmented across state records, social media aliases, and a handful of news articles from different decades. What followed was a process that most people don't realize exists: reconstructing a biography from public data that was never meant to be connected. If you are trying to find out who someone is beyond their professional output, here is how the actual work goes. I am going to walk through the method I used, the dead ends I hit, and the workaround that finally produced results. This is not about doxxing anyone. It is about understanding what public information reveals when you know where to look.
Unveiling Lorieal Sarkisian The Mysterious Woman Behind The Iconic Brand
The first step is always establishing what you already have. In my case, I had a name, a brand association, and roughly four years of timeline context. The brand itself had press coverage, interviews, and product launches going back to the early 2010s. What was missing was the person behind it. Most people assume that every notable figure has a Wikipedia page or a verified social media presence. That assumption is wrong for a significant number of founders, designers, and creative directors who operate below the threshold of mainstream visibility. I started with basic name searches, which sounds obvious but most people stop there. The trick is to modify the search strategy on the first iteration. If you type a full name into Google, you will get noise. The signal comes from narrowing. I searched the name combined with the company's incorporation state, then the city where the brand was first registered, then the executive title listed in the business filing. Each combination produced a different cluster of results. The business registration gave me a date of formation. The city record gave me a physical address that was later redacted but still showed in older archives. Here is the part nobody tells you about public records research: the data is not centralized. Every state maintains its own system, and the quality of digitization varies wildly. Delaware is well organized. Texas is a mess of scanned PDFs. California sits somewhere in between. When I hit the TexasSecretaryOfState website, I had to download over forty individual PDFs to piece together the ownership structure. Each one took about eight minutes to open and parse. I spent roughly five hours on that state alone before I found the name I was looking for nested inside an amendment filed in 2016.
The next layer is court records. I was not looking for lawsuits. I was looking for any mention of the name in civil filings, probate cases, or family court documents that happened to become part of the public record. This is where most people give up because the interface is terrible. Go to justia.com or courtlistener.com and search the name along with keywords from the brand. You will find things that were never meant to be associated with the company. A property transfer. A traffic infraction. A name change petition. These documents contain addresses, phone numbers, and sometimes dates of birth that you can cross-reference. I ran into a specific problem with Lorieal Sarkisian that took me two weeks to solve. The name appears in multiple geographic regions, which means standard search algorithms return conflicting results. When I searched for the name plus Texas, I got a hit for a different person who lived in Houston and shared the same name. When I searched for the name plus California, I got another unrelated person. The actual person I was researching lived in neither state at the time of my search, which is why every query was bouncing between false positives. The workaround was to use the brand's supply chain data instead. I pulled shipping manifests from customs records, which list the actual business address on file with CBP. That address pointed to a region in the Southeast that none of my earlier queries had targeted. Once I narrowed the geographic scope to that area, the name searches started returning consistent results. I cross-referenced voter registration records for that county, which gave me a date range. I then searched that date range against social media platforms using advanced operators that most people do not know exist.
Get the Full Details
Social media is both the most useful and the most deceptive source. People curate their profiles differently depending on the platform. LinkedIn shows professional history. Instagram shows visual clues like background locations, event check-ins, and clothing styles that can help date photographs. Twitter shows the timeline of when someone started paying attention to certain topics. The key is to treat each platform as a separate document rather than a unified identity. The person posting about fashion on Instagram may not be the same person filing business documents, even if the name matches. I encountered one counter-intuitive finding that changed the entire direction of the research. I had been operating under the assumption that Lorieal Sarkisian was the founder. The business records suggested something different. The name appeared as a registered agent, not an owner. A registered agent is a legal placeholder, not necessarily the person making decisions. This is a detail that almost everyone misses because the terminology sounds important but carries a completely different meaning in practice. The actual owners were listed elsewhere in the paperwork, under different names that did not match what I was searching for. This realization forced me to start over with a different hypothesis. Instead of searching for the founder, I searched for the design team. Product brands are built by teams, not individuals. The public face is often a marketing creation. By tracing the employees listed in patent filings, design registrations, and industry conference speakers, I found a smaller group of people who were consistently present across multiple projects. One of those people had a name that appeared in a completely different context: a university alumni publication from 2008. That publication contained a photograph and a short biography that mentioned the brand before it existed.
The timeline reconstruction took another ten days. I collected every date I could verify, then built a spreadsheet. Column A was the date. Column B was the source. Column C was the event. Column D was my confidence level on that entry. This is what separates amateurs from people who actually know what they are doing. Confidence levels matter because public records are frequently wrong or outdated. A business address can remain on file for years after someone has moved. A court document can list a stale phone number. Your job is to track which entries are solid and which are provisional. I also found that email addresses are more valuable than most people realize. When you search for a name along with the domain of the brand, you sometimes find forum posts, mailing list archives, or conference attendee lists that include the person's work email. Those emails follow patterns that reveal the company's naming convention. Once you know the pattern, you can generate likely addresses for other employees. This is useful for verifying whether a person you found on one platform is actually connected to the brand. There are limitations to this approach that I want to be blunt about. It does not work for everyone. Some people deliberately structure their businesses to minimize public exposure. They use LLCs layered inside other LLCs. They operate in states with weak disclosure laws. They have no social media presence at all. In those cases, you will hit a wall, and there is no workaround. The information simply does not exist in the public domain.
Even when the data is available, it can be misleading. A name match is not a person match. An address match is not a current residence match. A date range is not a birthdate. I learned this the hard way when I spent three days building a profile on someone who turned out to be a distant relative with the same name who had passed away in 2019. The obituary was the first thing I should have checked. Public records do not die with the person they describe. They persist in archived form, and searching for them without that awareness will produce inaccurate conclusions. The tools you need are mostly free. Wayback Machine for tracking when a website changed. Justia for court documents. State secretary of state websites for business filings. LexisNexis or similar services if you have access through a library or employer. Social media advanced search operators. A spreadsheet and a lot of patience. The total time investment for a thorough investigation of one person is typically two to four weeks if you are working part-time. Full-time researchers in this space complete similar investigations in one to two weeks. If you are starting this process, begin with the brand, not the person. Establish what the company does, where it is registered, who files its documents, and when it was founded. Then work backward from the corporate structure to the individuals. This reverses the intuitive approach most people take, but it is significantly more efficient because the business entities have paper trails while the people themselves often do not.
One more thing that took me a long time to understand: credibility assessment. Not everything you find is true. People lie on public forms. Addresses are intentionally incorrect sometimes. Names are misspelled in court records. Your job is not to accept what you find but to verify it through independent sources. If a business filing says one thing and a court document says another, you need at least one additional source before you can decide which is correct. Three sources minimum for anything you plan to share with other people. The research process changes how you see the internet. You start noticing patterns in how people present themselves across different platforms. You learn to distinguish between curated identity and accidental exposure. You develop a sense for which documents are reliable and which are garbage. Most importantly, you learn that mystery is usually just incomplete data, and incomplete data is solvable with the right method and enough time.
