About Novus Data
Who it is for
Physical trade breaks before prices move. Novus Data keeps a register of what is going wrong across the shipping lanes, ports, chokepoints and trade rules that carry the world’s goods — and maps each problem to the companies and sectors it reaches, with the mechanism and the source stated every time.
It is written for investors and analysts who need to know how a disruption in physical trade reaches prices, earnings and risk. It is also read by procurement and logistics managers who plan around freight cost and transit time, operators and founders whose landed costs move with shipping and anyone following trade policy closely enough to need the detail.
What it covers
Novus Data follows the physical and regulatory machinery of trade: shipping and freight, chokepoints, ports, trade policy, concentrated industrial inputs, the energy costs attached to moving goods, and the economic releases that describe all of it.
What is going wrong at any moment is kept in the register, and the exposure chart maps each entry to the companies and sectors it reaches. The Novus Data Briefing summarises the movement in both by email.
How it is produced
Everything starts from primary sources wherever they exist: canal and port authority notices, customs and trade statistics, regulatory texts and official releases, and the filings and announcements of the companies involved. Trade press and carrier commentary are used to find stories, not to settle them.
Every entry in the register carries the date it was last reviewed, and every company or sector named against a disruption carries three things: the mechanism by which the problem reaches it, how confident that assessment is, and at least one source you can follow. An assessment that cannot supply all three is not published — it is dropped by the site itself, not left to editorial discretion.
The work is reading rather than modelling. Novus Data does not run a proprietary dataset and does not publish forecasts dressed as numbers. Where something is uncertain, it is marked as inferred or estimated rather than stated flatly.
The standard is deliberately awkward to meet. It is easy to write that a company is "exposed" to a problem; it is much harder to say by what mechanism, how well established that is, and when it was last checked. Requiring all of it means the chart fills slowly — and that anything on it is worth the space it takes.
Novus Data publishes analysis and commentary. It is not investment advice, it is not a recommendation to buy or sell any security, and it is not a substitute for your own work.
What it takes to appear on the exposure chart
Naming a company against a problem is a claim someone may act on, so every one of them has to carry four things. An assessment missing any of them is not published — the site drops it when the page is built, rather than leaving it to editorial discretion.
- A mechanism. The sentence explaining how the disruption reaches that company. “Affected” is not a finding.
- A confidence level. How well established the assessment is, stated on the chart itself.
- A date. When the assessment was last true, shown on every cell and every entry.
- A source. At least one, with a publisher and a link you can follow.
What the confidence levels mean
- Reported
- Stated by the company, a regulator, or a named report.
- Inferred
- Follows from disclosed facts such as filings or published routings.
- Estimated
- A judgement from partial information. Treat as the weakest claim here.
Severity is a three-step scale — low, moderate, high — and no finer. A more granular scale would imply a precision that reading public sources cannot support.
How old is too old
Every register entry carries the date it was last reviewed. An entry not reviewed within 21 days says so on its own page rather than presenting itself as current. The review date is the figure to trust: it is a fact about when the work was done, and it does not go out of date the way a relative “three days ago” would.
Corrections
If something here is wrong, it gets corrected rather than quietly edited. A correction to a register entry is made on the entry itself, the review date is updated, and what changed is stated in the briefing that follows.
Assessments are withdrawn as readily as they are published. If the evidence behind an exposure stops holding, the exposure is removed from the chart — a claim is only as good as the source under it, and there is no benefit to defending one that has stopped being true.
Corrections are the most useful thing a reader can send, and they are read first.
Who writes it
Novus Data is written by Gavin McGreevy. I research and write every issue.
Contact
Corrections, questions and sources are welcome at aandg.novusdata@gmail.com. Corrections are the most useful thing you can send.
Subscribe
Delivered by email. Free.