Editorial standards

What we publish

Operator-grade Namibian commercial intelligence: the friction, the numbers, the named projects. Coverage is shaped by what the cross-border investor actually prices into a deal: peg risk, permit timelines, port congestion, empowerment math, repatriation latency.

How we verify

Every load-bearing fact is hyperlinked to a primary source. Where a primary source does not exist, we say so. Where we estimate, we publish the assumptions. If a number changes after publication, the article is updated with a dated correction at the top.

For numerical analysis (cap tables, IRR sensitivities, tracker readings) we publish the working. Members can request the underlying spreadsheet.

What we refuse

• Press-release journalism. If a release would have made it into 30 newsletters by 9am, we don't repeat it. We add the friction nobody else publishes.
• Sponsored editorial. Display ads, yes; pay-for-coverage, never.
• Boosterism. "Africa rising", "untapped potential", "the next [whatever]" don't appear in our copy. We assume the reader has capital at risk.
• Anonymous attacks. Critical pieces name the project, the operator, and the number. We do not run anonymously-sourced character work.

Corrections

We log every material correction at /corrections with the date, the original assertion, and the corrected fact. Typos and grammatical fixes are not logged.

Conflicts

Reporters disclose holdings in any named company before publication. If a piece touches an editor's prior employer, the byline says so. The masthead's annual disclosure runs at /disclosure.

AI use

We use language models for first-draft scaffolding, fact-checking, and translation. Every published article is read by a human editor before posting. The editorial responsibility for accuracy is the byline's; the publisher's responsibility for the editorial standard is ours.