A newsroom site is an archive with a front page

Search for how to build a news website in Kenya and the answers are about picking a theme, a host and an ad network. Those are real decisions and none of them is the one that decides whether the publication survives its own redesign.

A publication that has been running for years is carrying something that cannot be rebuilt: a set of addresses other sites have linked to, search engines have indexed, and readers have bookmarked. The design can be redone. The addresses cannot be recreated by effort. Any plan that treats them as cleanup after launch is a plan to make the publication smaller.

Before the design, enumerate every URL the current site serves. Not the ones in the navigation. Every one, pulled from the CMS, the server logs, the sitemap and Search Console, then merged, because each source knows about addresses the others do not.

Give each URL a decision: it moves, it stays, or it is retired with a reason written down. Then request every entry against the deployed site and read the response. A single permanent redirect, no chains, and a destination that returns 200. A map that has not been requested against production is a hypothesis. On the Business Report rebuild that run was the gate: a failure in it would have been a blocker, not a snag.

The failure mode of a news site is not design. It is a system where publishing anything slightly unusual, a correction, a live blog, an author who is also a source, an article that must be unpublished, requires a developer. Editors route around it, and within a year the structure has decayed into one free-text field.

So the content model comes first and it is built with the people who file copy. Sections, authors, categories and articles as real records rather than tags on a blob. Corrections and updates as first-class events. Preview that shows exactly what will publish. Structure an editor can work inside is structure that stays intact, and it is also what keeps an archive searchable a decade later.

A business publication is read the day it publishes and cited for years afterwards, and those two demands are not served by the same caching strategy. The front page needs to move the moment a story breaks. An article from three years ago can be built once and served from the edge indefinitely.

Both have to be readable on a phone over mobile data before any JavaScript arrives, which means server rendering rather than assembling the page in the browser. The next largest weight is usually images sent at several times the size they are displayed, and on a site with thousands of articles that is also the line on the hosting invoice that grows with the archive rather than with the audience.

A newsroom also needs the surfaces a reader never thinks about until they are missing. RSS and a sitemap generated from the same records as the site, so a section cannot be added and forgotten. On-site search over the whole archive. Social cards and newsletters wired to the article record rather than maintained separately. Stable URLs the archive can be cited by.

Then the handover. The newsroom publishes without the developer before the project is considered delivered. A news site that needs an engineer to post an ordinary story on a Sunday evening is not finished, however it looks.