Methodology
How ideas get made.
Most research shops describe their output. We describe the machine. This page is how an idea travels from raw signal to the note in a member’s inbox — the whole route, nothing held back.
[01] The pipeline
Machines go wide. We go deep.
The wide end is automated. Our own pipeline ingests filings, prices, and public data continuously, then scores and clusters what it finds — overlaid with the demographic and alternative-data signals we collect ourselves. Its job is not to conclude anything. Its job is to point.
The narrow end is human. From what the machine surfaces, we pick one idea and spend months on it — primary documents, the numbers rebuilt from scratch, the bear and bull case argued until one of them breaks. Most candidates die here. The ones that survive become a note, and the note goes to every member at the same moment. No tiers, no early access, no private preview.
[02] Anatomy of an idea
Every note carries the same skeleton.
- Where consensus sits, and why we think it’s wrong
- The asymmetry — what’s priced in versus what we expect
- What would prove us wrong, stated before it can
- A conclusion — act, watch for a trigger, or walk away
- Any position we hold, disclosed on page one
A note that ends in a pass is still a finished note. We publish the conclusion the research earns, not the one a publishing schedule demands.
[03] Publication rules
Impersonal. Simultaneous. Disclosed.
- Impersonal
- The same note goes to every member, and we advise no one individually. The reader owns the decision and the sizing — always.
- Simultaneous
- One send, all members at the same moment. Nobody reads it early, and that includes anyone who pays us.
- Disclosed
- If accounts we manage hold a position, page one says so — as of publication day. After that, silence by design: no ratings, no coverage universe, and never a buy or sell signal.
The full disclosure policy, and every recommendation from the past twelve months, sits at /disclosures.