Most of your traffic lands on something written years ago, from a search result, on a phone, with an ad stack loading on top of it.
Publishers are unusual in that the majority of value sits in content nobody is currently editing.
Every redesign, CMS move and category rename leaves a layer of redirects behind. Those are resolved at the edge and kept as a permanent record rather than as a plugin's database table.
Header bidding, consent and measurement scripts all compete with the article. Their weight is visible, so the revenue-versus-speed conversation is based on numbers rather than on instinct.
Feeds, AMP-style endpoints and partner exports are polled constantly by machines. They are cached and revalidated, so aggregators are not a second audience the origin has to render for.
Two hundred thousand articles with tags, authors and series is a large dataset. Archive and author pages are indexed rather than left to scan the post table.
Editing a decade-old piece should update one page, not invalidate the whole archive at the edge. Purges are scoped to what actually changed.
A newsroom's attention is on today. Its traffic is spread across everything published since the site launched, and that mismatch decides what the hosting has to be good at.
A publisher's Core Web Vitals are rarely decided by the server. They are decided by what loads on top of a page the server already delivered quickly.
Every row follows from the archive being both the asset and the traffic source.
| Setting | What we do | Why |
|---|---|---|
| Redirect history | Retained across every past move and resolved at the edge | A publisher accumulates layers of redirects from years of redesigns, and evaluating them in PHP on every miss makes the oldest URLs the slowest ones. |
| Third-party budget | Ad and measurement script weight reported | Server response is rarely what fails a publisher's Core Web Vitals, and without a number the revenue-versus-speed argument has nothing to stand on. |
| Feed delivery | Cached with validators for aggregators and partners | Syndication endpoints are polled by machines on a timer, which makes them more frequently requested than any article on the site. |
| Archive indexing | Author, tag and date queries indexed | Section and author pages scan a post table with hundreds of thousands of rows, and they are exactly where search traffic lands. |
| Purge scope | Narrowed to the content that changed | Correcting one old article should not invalidate an archive that took hours of traffic to warm. |
Every plan includes free migration, daily backups, SSL and 24/7 support.
Move the publication across — migration is free, redirects included.