# International SaaS SEO
> How to pick your first markets from real signup data, choose subfolders over ccTLDs, implement hreflang correctly, and localise pricing without splitting rankings.
Source: https://saas-marketing.net/guides/international-saas-seo/
Topic: SaaS SEO
Type: guide
Published: 2026-09-11
Last updated: 2026-09-11
Publisher: SaaS Marketing (saas-marketing.net)
License: CC BY 4.0. Quote or republish with attribution and a link to https://saas-marketing.net/guides/international-saas-seo/
## Short answer
Pick international markets from your own data: existing organic sessions, trial signups and win rate by country, not a TAM deck. Use subfolders on your main domain rather than ccTLDs, because a SaaS domain's authority is the asset you cannot rebuild. Implement hreflang with reciprocal return tags and an x-default. Translate documentation and support content first, since it compounds retention and rankings at the same time.
## Key takeaways
- Your existing country-level signup data is a better market selection input than any market sizing deck.
- Subfolders beat ccTLDs for almost every SaaS company, because you cannot rebuild domain authority in five markets at once.
- Four hreflang errors cause most failures: missing return tags, invalid codes, pointing at non-canonical URLs, and no x-default.
- Translate docs and support content first. It lifts retention and organic coverage from the same spend.
- Never machine translate a comparison page. Competitive claims need a native speaker who understands the local market.
- The Netherlands and the Nordics rank highest for English proficiency, so translating there often returns less than DACH or France.
---
Most SaaS international SEO projects start with a slide showing the European software market in billions of euros. That slide has never once predicted which language version would actually produce revenue. The teams that get this right start somewhere much duller: a country breakdown of trial signups they are already getting without trying.
Localisation is expensive, slow to reverse, and it permanently increases the maintenance cost of every page you own. So the first decision, which markets, is worth more than every technical decision that follows.
## Pick your first two markets from your own data, not a TAM deck
You already have the answer in your analytics. Pull four numbers by country and the shortlist builds itself: organic sessions, trial or demo signups, signup-to-paid conversion, and average contract value. Most B2B SaaS companies discover that two or three non-English markets are already producing unpaid signups at a rate that has nothing to do with any marketing they did.
Run the query over the last 12 months and build a table like this.
| Country | Organic sessions | Signups | Signup to paid | ACV index | Read |
| --- | --- | --- | --- | --- | --- |
| Germany | 41,200 | 690 | 4.1% | 1.15 | Strong latent demand, low conversion. Localise. |
| Netherlands | 18,600 | 410 | 9.2% | 0.95 | Converting fine in English. Defer translation. |
| France | 22,400 | 240 | 2.8% | 1.05 | Demand exists, conversion collapses. Localise. |
| Brazil | 33,000 | 520 | 1.4% | 0.40 | Volume without value. Watch, do not invest. |
| Japan | 4,100 | 60 | 11.0% | 1.60 | Tiny traffic, excellent economics. Investigate. |
The diagnostic is the gap between traffic and conversion. High sessions with low conversion means buyers find you and then hit a wall, usually language, currency or a missing local payment method. That gap is the localisation opportunity. High sessions with high conversion means the market is already working in English and translation will return less than you think, which is exactly the Netherlands and Nordics situation: the EF English Proficiency Index has ranked the Netherlands first or near first for years, and Dutch software buyers evaluate English documentation without friction.
Low traffic with strong conversion and high ACV, like the Japan row above, is the most interesting signal on the table and the one most teams ignore. A market where sixty signups convert at 11 percent is telling you something about fit that thirty-three thousand Brazilian sessions are not.
Before committing to a market, check support ticket language and sales call recordings. If your AEs are already handling German procurement questions and security reviews in German, the demand is more mature than the analytics suggest, and the pricing page is probably not the bottleneck.
Two markets. Not five. The failure mode I see most often is a company launching six languages in one quarter, maintaining none of them, and having eleven hundred stale translated URLs eighteen months later. Pick two, get one to work, then decide.
## Subfolder, subdomain or ccTLD: the decision most SaaS teams overthink
Use subfolders on your main domain. `example.com/de/` rather than `de.example.com` or `example.de`. For a SaaS company the domain is the asset, and every ccTLD you launch starts from zero referring domains in a market where you have no relationships and no press.
The ccTLD argument that does hold water: enterprise procurement in some markets, and regulated sectors in particular, treat a local domain as a signal of local presence. If you are selling compliance software to German Mittelstand buyers who need a German entity and a German DPA anyway, the ccTLD is part of a bigger commercial decision, not an SEO one. Make it with the finance and legal teams, not with your SEO consultant.
One thing that is not negotiable: never switch language based on IP address alone with no crawlable URL. Googlebot crawls predominantly from US IPs, so IP-based redirection means your German pages may never be seen. Offer a suggestion banner if you must, and let the user choose.
If you are moving from an existing structure, treat it as a migration with all the usual discipline. The redirect mapping, staging validation and monitoring cadence in the [SaaS website migration SEO](/playbooks/saas-website-migration-seo/) playbook applies directly, and a botched language rollout can cost you rankings in the market you already had.
## Hreflang, and the four errors that cause most failures
Hreflang tells search engines that two URLs are the same content for different audiences. It does not boost rankings. It stops the wrong language version outranking the right one and stops your own pages competing with each other.
```html
```
Every page in that cluster carries the identical block, including a self-referencing tag. That reciprocity is the whole mechanism, and it's where implementations break.
The four errors, in the order I find them:
1. **Missing return tags.** The English page links to German, the German page does not link back. Google ignores the relationship entirely, so a one-directional implementation is the same as no implementation. This is by far the most common bug and it usually appears when one language is managed in a different CMS.
2. **Invalid codes.** `en-UK` is not valid, the country code is `GB`. Language is ISO 639-1, region is ISO 3166-1 Alpha 2, and region without language (`hreflang="de-AT"` is fine, `hreflang="AT"` is not) is invalid. Codes are case-insensitive but the order is not.
3. **Pointing at non-canonical or redirecting URLs.** If `/de/preise` 301s to `/de/preise/`, the hreflang cluster is broken. Every target must be a self-canonical 200.
4. **No x-default.** Without it, a user in a market you have not localised gets an arbitrary version. Point x-default at your primary English page.
Search Console retired its International Targeting report in 2022, so validation now happens in a crawler. Screaming Frog, Sitebulb, Ahrefs Site Audit and Semrush all report hreflang reciprocity errors, and the differences between them matter less than running one of them on a schedule. Our [Ahrefs vs Semrush](/comparisons/ahrefs-vs-semrush/) comparison covers which handles large multilingual crawls better if you are choosing now.
Sitemap-level hreflang scales better for large sites and avoids bloating the head on templated pages. Implementing both and letting them disagree, which happens the moment two teams own two systems, produces conflicting signals and a support ticket nobody can close.
## Translate, transcreate, or leave it in English, by page type
Not every page deserves the same treatment, and treating them uniformly is how localisation budgets get wasted. Sort by whether the content is factual, persuasive or competitive.
| Page type | Treatment | Why |
| --- | --- | --- |
| Product documentation, API reference | Translate, machine translation with human review is fine | Factual, high volume, heavily searched, compounds retention |
| Help centre and troubleshooting | Translate | Deflects support cost in the same spend |
| Feature and product pages | Translate with a native editor pass | Factual but persuasive. Needs local phrasing |
| Pricing | Transcreate plus local currency and tax handling | Highest intent page in any language |
| Comparison and alternatives pages | Transcreate, never machine translate | Competitive set differs by market. Claims carry legal risk |
| Case studies | Rewrite with local customers where possible | A US logo persuades a German buyer far less than a German one |
| Blog and top of funnel | Selective. Translate the winners only | Most posts do not earn their translation cost |
| Careers, legal, trust pages | Translate where required by law | Compliance driven, not SEO driven |
Translate the documentation first. That is the strongest opinion in this guide and the one most often argued with. Docs and help content are usually the largest indexable surface a SaaS company owns, they attract long tail queries in every language, and translating them reduces support load and improves activation in the same market you are trying to win. You get retention and rankings from one budget line. Marketing pages give you only the second.
Comparison pages are the opposite case. A German buyer's alternatives to your product are frequently different from an American's, local competitors exist that your English page never mentions, and a competitive claim that is fine in US comparative advertising can be actionable in Germany, where competitor comparison rules are stricter. Machine translating those pages is the fastest way to turn an SEO asset into a legal problem. Rebuild them with a native speaker who knows the market, using the same structure as your English versions from the [bottom of funnel page sprint](/courses/saas-seo-sprint/02-bofu-pages/).
## Keyword research in a market where you cannot read the results
Direct translation of your English keyword map produces a map of phrases nobody searches. German software buyers frequently search the English product category term alongside German modifiers, French buyers often do not, and Spanish splits hard between Spain and LATAM usage.
Work in this order. Take your existing [SaaS keyword map](/templates/saas-keyword-map-template/), strip it to the intents rather than the strings, then rebuild the strings per market with a native speaker looking at live results. Check three things for every core term: whether the local market uses the English loanword, whether the SERP returns vendors or dictionaries, and which competitors appear that you have never heard of.
That last check matters commercially, not only for SEO. Discovering that a local vendor owns the category term in France is a product marketing finding, and it usually changes the positioning before it changes the keyword list. Feed the result into the same brief structure you use at home, whether that is editorial or the [programmatic page brief template](/templates/programmatic-page-brief-template/) for integration and use case sets.
## Localising pricing, currency and payment
The pricing page carries the highest intent traffic in every language, so localise it properly or do not bother with the rest.
- Show local currency with a real price, not a converted estimate that changes daily. Stripe and Paddle both support per-market price points, and Paddle as merchant of record removes a large amount of VAT and invoicing work for teams without a finance function.
- Display tax handling the way the market expects. EU buyers expect to see whether VAT is included and B2B buyers expect a reverse charge option with a VAT number field.
- Support local payment methods: SEPA direct debit and invoice with purchase order in Germany, iDEAL in the Netherlands, Bancontact in Belgium, Boleto and Pix in Brazil. Card-only checkout quietly kills conversion in several European markets.
- Offer annual invoicing in DACH. German mid-market buyers frequently prefer invoice terms to card payment, and the absence of it stalls deals that looked ready.
Expect the sales motion to change too. Deal cycles in DACH and Japan typically run longer than US equivalents because of procurement and security review depth, and CAC usually rises in the first two or three quarters before local organic and word of mouth compound. Plan the payback window accordingly rather than judging a new market on a single quarter.
## The sequence most B2B SaaS companies follow, and why
**A realistic international rollout order**
That last step is the one nobody schedules and everyone needs. Set the review date when you launch the market, and write down the number that has to be true.
## What it costs, and the honest failure mode
Budget for a first market at roughly 15,000 to 50,000 dollars in year one for a 40-page marketing set, the documentation, and the pricing and checkout work. Professional human translation runs about 0.10 to 0.25 dollars per word for B2B software content, machine translation with post-editing roughly half that, and native transcreation of comparison and landing pages is priced closer to copywriting than translation.
Then the part nobody budgets: maintenance. Every product launch, pricing change and repositioning now has to propagate across every language. A company shipping weekly will find that translated pages drift out of date within two quarters unless localisation is wired into the release process rather than run as a project. Stale German pricing is worse than no German pricing.
The honest failure mode is this. Companies localise the marketing site, leave the product, onboarding emails and support in English, and then wonder why German trial-to-paid conversion barely moves. Buyers convert on the whole experience. If the product is English-only, translating the blog changes very little, and that is a product roadmap conversation before it is an SEO one. It is also the reason an early-stage company is usually better served by the domestic work in [seed stage SaaS SEO](/playbooks/seed-stage-saas-seo/) than by a second language.
## Start here
Pull the country table this week. If one market shows meaningful organic sessions and materially worse conversion than your home market, you have found your first localisation target and you did not need a market sizing exercise to do it. Scope that single market end to end, including the product and support surfaces, and give it four quarters against a number you wrote down in advance.
Then wire it into the plan rather than running it beside one. The sequencing in the [90 day SaaS SEO plan](/playbooks/saas-seo-90-day-plan/) and the market entry commercial work in the [SaaS international expansion playbook](/playbooks/saas-international-expansion/) both assume localisation is one workstream inside a programme, which is the only way it survives a busy quarter. For the link side, local coverage and local citations behave differently from your home market, and the tactics in [earn links and AI citations](/courses/saas-seo-sprint/04-links-and-citations/) need a native speaker to run.
## Frequently asked questions
### Should a SaaS company use subfolders, subdomains or ccTLDs for international SEO?
Subfolders, in almost every case. A SaaS company's domain authority is concentrated in one property, and splitting it across five ccTLDs means starting authority from zero in each market. Use ccTLDs only when you have a genuine legal entity, local pricing and a local team per market, or where a regulator or buyer expectation requires a local domain.
### Which markets should a B2B SaaS company expand to first?
Look at where you already get unpaid signups. For most English-first B2B SaaS companies the common sequence is DACH, then France, then Benelux and the Nordics, then LATAM or Japan. DACH usually justifies full localisation first because German-language buying behaviour is strongest, while the Nordics and Netherlands often convert fine in English.
### What are the most common hreflang mistakes?
Four account for most failures: missing reciprocal return tags between language versions, invalid codes such as en-UK instead of en-GB, hreflang pointing at URLs that redirect or are not self-canonical, and no x-default for the fallback page. Google ignores the entire cluster when the return tags are not reciprocal, so a one-sided implementation does nothing.
### Is machine translation bad for SEO?
Google's guidance targets low-quality automated content, not translation as a technique. Machine translation with genuine human review is acceptable and widely used. Raw unedited output is risky for marketing pages and actively dangerous on comparison pages, where a mistranslated competitive claim becomes a legal and credibility problem, not only an SEO one.
### What is the difference between translation and transcreation?
Translation preserves meaning sentence by sentence and suits documentation, API references and help articles. Transcreation rewrites for the market: different examples, different competitors, different objections, sometimes a different argument. Comparison pages, landing pages and case studies need transcreation, because the competitive set and the buying objections differ by country.
### Do you need local pricing pages for international SEO?
Local currency and local payment methods affect conversion more than rankings, but the pricing page is usually the highest intent page in any language. Show local currency, display VAT handling clearly for EU buyers, and support the payment methods that market expects, such as SEPA direct debit in Germany and iDEAL in the Netherlands.
### How much does it cost to localise a SaaS site for one market?
Professional human translation typically runs 0.10 to 0.25 US dollars per word, with machine translation plus human post-editing at roughly half that. A first market covering 40 core pages, the docs set and the pricing page commonly lands between 15,000 and 50,000 dollars in year one, before the ongoing cost of keeping it current as the product changes.