Why US SEO starts from zero
At home you sell in a market that knows you. There is a name, there are referrals, there are rankings built over years. In a new market you have none of the three.
Your domain carries no history there. The competitors are different, and most of them have been in that market longer than you. And what people type into Google is not a translation of what they type at home.
That is why most market entries fail quietly. Not because the content is bad, but because it was built for another market and translated.
If you are coming from Israel
An Israeli company entering the US market usually arrives with a site that already ranks in Israel, a product that already sells, and the assumption that the English version is most of the work. It is not. Three things change.
The buyer searches differently. Israeli tech buyers already search in English, so the Hebrew site taught you very little about American ones. US search intent is longer, more comparative, and shaped by competitors you have not met yet. The keyword list starts over.
Authority has to be rebuilt in the market. Mentions from Israeli press, Israeli directories and Israeli partners carry almost no weight for a US query. Google and the AI engines look for signals from inside the market: US publications, US customers, US profiles.
One person has to own the whole chain. Market entry is where an agency for content, a freelancer for the site and a consultant for SEO fall out of sync. I plan the strategy, write the content, build the site and run the SEO, so the US version is one decision, not three handoffs.
I have worked this from both directions: an Israeli company selling to the US market, a Canadian practice that went from 5 leads a month to a peak of 48, and one brand across six markets at once.
What changes when you add a market
Adding a market to a site that already ranks at home is a different job from launching a new site. The home rankings are the asset, and the first decision is how not to damage them.
Where the new market lives (a folder on the same domain, a subdomain or a new domain), how Google is told which version belongs to which country, and which existing pages now have a twin that competes with them. Get this wrong and the new market cannibalises the old one before it has earned anything of its own.
This is the part worth deciding before a single page is written. It is also the part most teams decide by default, by letting the CMS choose.
Research for each market separately
A keyword list translated from the home market is the most common mistake and the most expensive one. The volumes are different, the intent is different, and the competitors are different. Each market gets its own research: what people search for, how they phrase it, who already ranks and how strong they are. That research is the first step of an international SEO strategy, and it decides which pages get built at all.
Structure and hreflang
A country domain, a subdomain or a folder. Each has a price, and the choice is hard to reverse once the pages exist. hreflang tells Google which version belongs to which market, and it breaks in predictable ways: wrong codes, missing return links, pages that point at each other but not at themselves. I wrote up the four ways hreflang breaks because every migration I have audited had at least one of them.
Localization, not translation
Translate what explains. Rewrite what sells. Policy pages, specifications and technical explanations survive translation. The home page, product pages and landing pages are rewritten, because they answer a question the buyer in that market asks, in the words that buyer uses.
Authority in the target market
Links and mentions from the home market do little for a query in the new one. Authority is rebuilt where the buyer is: publications, directories and profiles the market already trusts. This is slower than buying links and it is the only version that still works after the next update.
Measurement per market
One Search Console view for the whole site hides what matters. Each market is measured separately, because one of them almost always pulls the average up and hides the rest. Impressions, clicks and leads per market, every month, so a market that is not working is visible in month two, not month eight.
What this looks like in practice
Six markets at once for one brand, from Scandinavia to Latin America: Sweden, South Africa, Kuwait, the UAE, Korea and Mexico. Six separate strategies, not one playbook translated six times. An external agency I briefed produced in five languages, 15 to 20 pages a month, and every page went through the development team before it went live. I built an hreflang setup that made the six versions of the same page help each other instead of competing.
And from the other direction: an Israeli company selling to the US market, and a law practice in Canada. The order of decisions is the same every time, and I set it out in five decisions to make before entering a new market.
What is included and what is not
Included. Research for each market, the structure and hreflang decision, mapping pages to markets, localization of the content or direction for the team that does it, the technical work, and separate measurement per market.
Not included. Machine translation that reads like machine translation, bought links, and a promise to rank by a date in a market you do not know yet.
Translate the site or rewrite it for each market?
Translate what explains and rewrite what sells. Policy pages, specifications and technical explanations survive translation. The home page, product pages and landing pages are rewritten, because they answer a question that is asked differently in each market.
A separate domain for each country or a folder on the same site?
A folder on the same site in most cases, because it inherits the authority the site already has. A separate domain when the market needs a local brand, local hosting rules apply, or the business in that country runs separately. The decision is made once, before the first page, because reversing it is a migration.