How Welsh and Gaelic keyword research works in mainstream SEO tools

What to take away

Why bilingual keyword research breaks mainstream SEO tools

Welsh is an official language in Wales with a legal footing. Gaelic has official status in Scotland under the Gaelic Language (Scotland) Act 2005. Public bodies in both nations publish bilingually, and that obligation creates search demand that mainstream tools rarely model.

Most keyword databases are built from English language corpora. A Welsh phrase such as addysg cyfrwng Cymraeg or a Gaelic phrase such as foghlam Gàidhlig may return no volume, a null result, or a volume borrowed from a similar English term. None of those outcomes is safe to plan a campaign on.

Language detection is the first failure point. Tools guess the language of a query and then apply the wrong stemming rules. Welsh morphology mutates initial consonants, so a word changes shape by grammatical context. Gaelic adds lenition and case endings. English stemming logic has no answer for either.

Source coverage is the second. Welsh language pages are a small share of the indexed web, and Gaelic pages are smaller still. A crawler that samples the web proportionally will under sample both. The result is a keyword tool that looks empty for a language that real people search in every day.

That gap is not a reason to skip the work. It is a reason to treat the tool as one input. Moz maintains a set of international SEO resources for language and region targeting that covers hreflang and locale structure.

That matters once you have a Welsh or Gaelic keyword list to place on the site.

ONS internet access data for Welsh and Scottish online usage context shows how thoroughly both nations are online, which is the demand side of the argument. Coverage is not the problem. Indexing is.

If you are choosing tools for a public sector or agency team in Wales or Scotland, the same question comes up in every procurement: what verified audience data sits behind the volume figure? For Welsh and Gaelic, the honest answer is often none.

Welsh keyword research workflow step by step

Welsh keyword research starts with the language people actually use in Wales, not a translation of your English list. Translated keywords follow English word order and miss the phrasing Welsh speakers type.

Step 3 is where teams go wrong. A zero volume figure is not proof of no demand. It is proof that the tool cannot see the demand. Record it and move on.

Use the phrase pairs to build a bilingual content map. One Welsh page and one English page for the same service, linked with hreflang, will outperform a single page with a language toggle. That is the core of the bilingual keyword workflow.

If your team is sizing the Welsh opportunity against other work, the split between market size and team capacity is covered in this audience research guide, and the logic holds for a Welsh language sub market.

Gaelic keyword research workflow step by step

Gaelic keyword research in Scotland is a smaller exercise with a different shape. The speaker base is concentrated in the Highlands and Islands, the publishing base is thin, and community and education queries dominate.

Step 5 matters more in Scotland than in Wales. A parent looking for Gaelic medium education may search in English, while a Gaelic speaker looking for the same service searches in Gaelic. Both are valid entry points and both need a landing page.

Community organisations are the best keyword source in Gaelic. Their newsletters, event pages and funding notices use current phrasing that no keyword database will surface.

Where you need to justify the spend to a board, the standard for competitor research that stands up applies here too: a named source for every claim, including the claim that a Gaelic keyword has demand.

Which SEO software handles Welsh and which handles Gaelic

The honest summary is that no mainstream tool handles either language well out of the box. Support varies by database, interface language and stemming. Here is how the common options behave.

Semrush is the most usable of the paid tools for this work because you can import a list and get competitor and SERP context around it. Its keyword research features tested against Welsh and Gaelic terms will return results for many terms, and blanks for others, which is itself informative.

The newer prompt research for AI-era keyword discovery in minority languages is worth testing. Prompt style queries behave differently from classic keywords, and Welsh and Gaelic users phrase questions in ways that suit that model.

Moz is not a Welsh or Gaelic database. Its SEO learning centre as foundational keyword research reference is still the right place to learn the method before you apply it to a minority language.

Ahrefs and Google Keyword Planner fill gaps. Neither is a Welsh or Gaelic tool, and neither should be the sole source for a bilingual brief.

Before you buy, decide which evidence should I check for language support claims. Vendor pages rarely state language coverage explicitly. Ask for a trial with your own Welsh or Gaelic seed list.

Diacritics, stemming and search volume gaps in both languages

Diacritics are the first practical problem. Gaelic uses the grave accent, as in Màiri or òran. Welsh uses the circumflex, as in tŷ, gŵyr and dŵr. Tools that strip accents will merge distinct terms or miss them entirely.

Test this yourself. Search a Welsh term with and without its circumflex in your chosen tool and compare the results. If the two queries return different volumes, the tool is treating them as separate words, which splits your data.

Stemming is the second problem. Welsh mutates word beginnings, so a term can appear as cath, gath or nghath depending on grammar. Gaelic applies lenition and case endings in the same way. English stemmers do not model either, so related queries are not grouped.

The fix is manual. Build your own grouping in a spreadsheet: one row per concept, one column per surface form. Then import the grouped list into the tool for volume and SERP checks.

Search volume gaps are the third problem. For most Welsh and Gaelic terms, no tool holds enough data to publish a volume. Where a figure does appear, treat it as indicative and validate against Search Console impressions from your own pages.

This is where a product comparison needs a named source for every volume claim. If a vendor cannot name the corpus behind a Welsh or Gaelic volume, do not put it in a business case.

Testing bilingual keyword sets before you commit budget

Do not commission a Welsh or Gaelic content programme on tool data alone. Test first, cheaply, and let the market answer. In Wales and Scotland alike, the test costs less than one mistargeted content brief.

A worked example makes the decision concrete. Suppose a Welsh public body has ten candidate Welsh terms. Paid search shows three with meaningful impressions and clicks, four with impressions but no clicks, and three with nothing. The three with clicks get pages first. The four with impressions get title and meta work. The three with nothing wait.

The same logic applies in Gaelic with smaller numbers. Thin data is still data if you collect it deliberately and review it on a fixed date.

Keep the test honest. Record the source of every volume figure, the date you checked it, and the tool version. A bilingual keyword workflow is only as good as its audit trail.

Common questions