A Spanish-language site rarely has a Spanish audience. It has a Spanish one and a Latin American one, ten thousand kilometres apart, reading the same pages.
The defining problem here is geographic, not linguistic: your readers are not in one place and no single origin is near all of them.
Static and full-page cache is answered from the nearest edge on either continent, which is what stops half your readership paying for the distance to the other half.
Signed-in and transactional pages always reach the origin, so its placement is a deliberate choice about which audience gets the shorter trip.
If you publish regional variants, each needs its own URL and language declaration. Without that, a Mexican reader gets the Spanish page and search engines see duplication.
The Spanish regulator asks where data is held and who processes it, including consent handling. Those are answered with a location and a list.
Spanish accents and ñ have to match whether or not the reader typed them. Collation is set so a search for 'diseno' still finds 'diseño'.
There is no origin location that is near both Madrid and Bogotá. The work is deciding which half pays for it, and making sure the other half only pays on uncached requests.
The honest measurement is the far side. Any configuration looks good from wherever the origin happens to be.
The audience is split across an ocean, which is a geography problem before it is anything else.
| Setting | What we do | Why |
|---|---|---|
| Edge coverage | Cached content answered on both continents | A Spanish-language audience is split between Iberia and Latin America, and without edge coverage one of those halves pays for the distance on every page. |
| Origin placement | Chosen explicitly, knowing it favours one side | No single origin is near both Madrid and Bogotá, so pretending the choice does not matter simply means it was made by default. |
| Regional variants | Given their own URLs and language declarations | Spanish differs enough between regions that many sites publish variants, and undeclared ones read to a crawler as duplicate pages. |
| Search collation | Accent-insensitive matching for Spanish characters | Readers omit accents when typing, so a strict collation makes a search for the correct word fail against the correctly spelled content. |
| Consent and data records | Prepared for AEPD questions specifically | The Spanish regulator asks about consent handling as well as storage, and those are separate answers that both need to exist in advance. |
Every plan includes free migration, daily backups, SSL and 24/7 support.
Move the site across — migration is free, placement discussed first.