Separate country domain or just a folder on the main site — which one actually moved first for you

My friends, i have doubt and everybody give me different answer.

I have a money site in english that work ok. Now i want the same for Brazil, all content in portuguese, and the question is do i buy the country domain or do i put a folder on the site i already have.

Everybody who write about this is from US or UK and they say folder because of the authority. But actually i never see one of these people rank anything in Brazil. And here the country domain is what people trust, is not only a search engine question, is a conversion question too (confiança - trust).

So i want real answer from somebody who did it. Not theory. Which one moved first for you and how long.

I've done both, on the same business, four years apart, which is as close to a comparison as I have and it is not very close.

The case I can actually report

A German client wanted Austria and Switzerland. In 2021 we did folders. In 2025 the same business wanted Poland and we did a separate country domain, mostly because the local team insisted for the same trust reason you're describing.

	Folders (2021)	Country domain (2025)
First page indexed	3 days	16 days
First non-brand ranking in top 20	~5 weeks	~14 weeks
Reached commercially useful volume	~4 months	~9 months
Conversion rate vs main site	0.7x	1.4x

The folder ranked faster. That part is not really in dispute and the mechanism is boring: it inherits crawl frequency and some site-level signals immediately. The country domain started from nothing, the way any new domain does.

Why the last row matters more than the first three

The country domain converted twice as well relative to the folder. If your Brazilian traffic converts at 1.4x on a local domain, nine months to volume at 1.4x beats four months at 0.7x within the first year and the gap widens after that.

So the honest answer to "which moved first" is folders, and the honest answer to "which was worth it" was the country domain, and those are different questions.

What would change my answer

1. If the local domain is expensive or restricted to register — some countries require a local presence, and paying an agent to hold a domain for you is an ongoing dependency I'd avoid.
2. If the content is a translation rather than written locally. A translated folder does badly and a translated country domain does worse, because you've paid the startup cost for a site that isn't actually local.
3. If you can't sustain it for nine months. The folder gets you something in month two. If your business needs revenue in month three, that decides it and nothing else matters.

Uncertain

I have no data at all on Brazil specifically and the country-domain trust effect is very different market to market. I'd weight local reports over mine.

Summary: folder ranks sooner, country domain converts better, pick based on how long you can wait.

nine months against four. most people here don't have a business that exists in nine months

Conversion row is the only row. Everything above it is a means to it.

Depends what you're running, and there's a middle option nobody has mentioned that I'd genuinely consider here.

Register the country domain now, park nothing on it, and point it at the folder. You've secured the name — which matters, because someone will register it once you start ranking, and buying it back off them costs more than the nine months — and you can switch the canonical relationship later once the folder is generating enough to justify the migration.

The migration is a real cost and I don't want to pretend otherwise. But it's a known cost you pay once, against an unknown cost of discovering in year two that the name you needed is gone.

APAC rather than LATAM, but the trust effect is the same and stronger in some markets than anything I've seen written about.

We ran a folder and a country domain in parallel for the same product in one market for five months. Folder had roughly three times the sessions. The country domain produced more revenue by month four.

One caveat that cost us time: if you go with the country domain, do not run both with the same content and hreflang between them as a hedge. We tried that and spent two months untangling which one was ranking for what. Pick one for each market and commit.

The folder advantage is smaller than it was and it's been shrinking for years, which nobody accounts for because everybody's mental model is from when they last did it.

In 2014 a folder on an established domain would rank in a new market in a fortnight, more or less regardless of what was in it. That's what people mean when they say folders inherit authority, and it's a memory of a specific era rather than a rule. By 2019 it was slower. Hexforge's 2021 figure of five weeks would have been a bad result in 2015.

What that means practically: the 2021 column of his table is probably optimistic for 2025 and the gap between the two options is likely narrower now than it looks there. Which pushes towards the country domain if you were already leaning that way for conversion reasons.

FWIW if you take byte_broker's middle route, redirect the country domain at the folder rather than serving the same content on both. Two live copies is how the hreflang mess hana_v described starts.

Seven weeks later, small update.

I did the middle thing. Country domain is registered and redirect to the folder like PipelineRat say. Folder is live since the 1st of September with fourteen pages written by a Brazilian person, not translation, because hexforge point 2 scared me.

Indexed in four days. First ranking that is not my brand at about week four, position 17 for something nobody search but is real. So the folder is fast, this part is confirmed.

The conversion question i cannot answer for maybe a year. Actually i think everybody in this thread agree the folder rank faster and nobody agree what happen after, so i will come back and say.