Research Methodology
Hynexly is a source-first investing publication. This page documents how the desk actually researches a company or market: which sources we trust, the order we read them in, how we structure a thesis, and what we never do. It is the standing reference behind every article's research-process desk note.
Last reviewed: June 9, 2026
Evidence-first sourcing
Every load-bearing fact maps to a source, and we work down a fixed hierarchy: primary filings (10-K, 10-Q, S-1), regulator and government data (SEC EDGAR, FINRA, EIA, central banks), company releases and investor-relations material, established market-data providers, then major financial reporting. A weakly sourced number never anchors a thesis; if the evidence is thin, the claim is marked unresolved rather than asserted.
Firsthand research process
We open the primary document before the headline. The desk note at the top of each analysis post records that process honestly — which filing we pulled first, which number reset our read, which popular framing we checked and could not support from the evidence. It is process narration grounded in the sources already cited in the piece, never opinion and never an invented anecdote.
How a thesis is structured
Company and stock posts carry an analyst-grade shape: a falsifiable thesis, a Source Evidence Snapshot built from captured primary figures, a frank read of what the market is pricing, concrete risks each paired with a confirming signal, and the specific upcoming event that would flip the call. The structure is fixed in purpose but varied in surface so the library does not read as one repeated template.
Source captures and evidence panels
When a source supports it, we show the real figure — a filing line, an exchange quote, a regulator chart — captured and rendered in a consistent evidence panel with a caption that states what the image shows and what the reader should notice. Captures are tightly cropped without hiding labels or context. We do not use AI-generated charts, decorative stock imagery, or creator-video screenshots as evidence.
What we never do
We do not invent analyst quotes or research-house names, fabricate DCFs or price targets, or issue BUY / SELL / HOLD ratings — we are not licensed, and a transparent framework serves readers better than a rating they cannot underwrite. We do not pad evidence sections with AI-generated explainer cards. A valuation range derived transparently from stated assumptions is acceptable; a single number presented as a target is not.
Updates and versioning
Investing facts age. When a key number or framing changes, we update the article rather than leaving a stale page in discovery, and trust pages like this one carry a dated review line. Strong existing pages are refreshed with new evidence before a near-duplicate is published.
Automation and the human gate
Automation may help draft and organize, but publication is gated by usefulness, not volume. A draft that is repetitive, too short, weakly sourced, or templated is blocked before it reaches discovery. Automated drafts pass the same source-grounding and editorial checks as everything else, and an editor — not the automation — makes the decision to publish.