Strategy
Part of Planning multilingual social marketing around self-described needs and reversible decisions
Census 2021 language figures for multilingual social marketing: counts, English proficiency and a shortlist
Main-language counts for England and Wales, English proficiency, Welsh as a separate case and local patterns, with a worksheet for shortlisting social languages.
English as a main language: the headline figure
Across England and Wales together, 91.1% (52.6 million) of usual residents aged three years and over had English (English or Welsh in Wales) as a main language in 2021. That was down from 92.3% (49.8 million) in 2011. Keep this combined geography attached to both figures.
The ONS main-language definition is a person's first or preferred language. Census estimates describe residents on Census Day, 21 March 2021. Keep demand signals separate from this population measure. Do not read a resident count as a total of followers, customers or requests for translated social posts.
A shortlisting worksheet
Polish, Romanian, Panjabi and Urdu were the most common main languages other than English (English or Welsh in Wales). Their counts and percentages cover England and Wales together. Use these names as starting candidates, not as a ranking of demand for your brand's social content.
| Language | Geography and measure | Census figure | Question before choosing |
|---|---|---|---|
| Polish | England and Wales together, main language | 612,000; 1.1% | Is it relevant to your target area? |
| Romanian | England and Wales together, main language | 472,000; 0.8% | Do intended readers prefer this content in Romanian? |
| Panjabi | England and Wales together, main language | 291,000; 0.5% | Do intended readers prefer this content in Panjabi? |
| Urdu | England and Wales together, main language | 270,000; 0.5% | Do intended readers prefer this content in Urdu? |
| Welsh | Wales only, able to speak Welsh, aged three years and over | 538,000; 17.8% | Is Welsh relevant to the content and place you serve? |
| Welsh | England only, main language | 7,000 | Have you checked content preference separately from main language? |
Romanian moved into the top 10 main languages, excluding English (English or Welsh in Wales), in 2021. Keep the Welsh entries separate from that ranking. The Wales row measures speaking ability; the England row measures main language. Neither is a like-for-like alternative to the combined main-language figures.
Read English proficiency separately
English language proficiency means how well people whose main language is not English (English or Welsh in Wales) speak English. In the census, it relates to how well a person feels they speak English. Read that measure separately from their first or preferred language.
People who did not report English as a main language were asked how well they could speak it: 8.9% (5.1 million) of residents. Of those 5.1 million, 43.9% (2.3 million) could speak English very well and 35.8% (1.8 million) well.
A further 17.1% (880,000) of those people could not speak English well, and 3.1% (161,000) could not speak it at all. Both sets of figures cover England and Wales together.
Across England and Wales together in 2021, a further 7.1% (4.1 million) of the overall population were proficient in English (English or Welsh in Wales) but did not speak it as their main language. For multilingual audience research, ask about preferred content language separately from English-speaking ability.
Local patterns can change the shortlist
The Bath and North East Somerset 2021 Census top-10 table gives 1,400 for Polish as a main language. Treat that as a local count, not a portion of your social audience. For a locally focused brand, check the relevant area before carrying over the combined ranking.
In Hammersmith and Fulham, 8 in 10 residents aged three years and over had English as a main language. French, Spanish and Italian were the other most common main languages. Add those names to a local shortlist rather than limiting it to the combined list.
In Westminster, 26% of residents do not have English as their main language. Arabic is the most commonly spoken language after English, but it differs in all wards. In the 2021 Census, 4% of residents reported that they could not speak English well or at all. Check ward relevance rather than assuming a borough-wide choice.
Keep Welsh as a separate case
In Wales, an estimated 538,000 usual residents aged three years and over (17.8%) reported being able to speak Welsh in 2021, down from 562,000 (19.0%) in 2011. This is a Wales-only ability measure. Do not position it as another England and Wales main-language percentage.
In Wales, the main-language question asked whether people's main language was anything other than English or Welsh. It cannot determine how many considered Welsh their main language. The Welsh ability question was not asked in England, where 7,000 people said Welsh was their main language in 2021.
The England main-language measure is likely not to capture everyone there who can speak Welsh. Welsh skills are self-assessed, and the Welsh census questions do not ask how often people speak Welsh or how well they speak the language. For Welsh and English localisation, distinguish ability from content preference.
The percentage of usual residents aged three years and over able to speak Welsh fell between 2011 and 2021 in all Welsh local authorities except Cardiff, Vale of Glamorgan, Rhondda Cynon Taf and Merthyr Tydfil. These exceptions describe change, not a ranking of language need. Check the place you serve rather than choosing from the trend alone.
From worksheet to working shortlist
Use the worksheet to narrow the question, not to estimate reach. Keep each candidate's geography and measure visible as you consider the content you want to publish. A shortlist should rest on relevance to intended readers, rather than the largest combined census count alone.
- Match the geography.Name the area your social content should serve. Add locally prominent languages that sit outside the combined ranking. Put the geography and candidate languages into your social media brief. Keep each resident count attached to its area, rather than presenting it as an audience forecast.
- Ask about the content.Ask intended readers which language they would prefer for the information you plan to publish. Keep that answer apart from main language and English proficiency. Use it to inform your channel strategy, without assuming that language preference establishes which platform someone uses.
- Record the shortlist.Illustrative example: "Polish stays on our shortlist because our intended readers prefer it for this information." Use that wording only when responses support it. Carry each shortlisted language into multilingual hashtag research, and keep census counts out of social reach claims.
