¿Hablas inglés? – the trouble with determining language

This post is going to be the first of many talking about geo location and internationalisation* of search results. To start I thought we would take a look at the problems the search engines have trying to ensure they return the most relevant results for any search you throw at them.
* My spell check just pointed out that I should spell internationalisation with a z, it made me giggle. Its like that ‘joke’ about why dyslexia is so hard to spell
Living and working in the UK and speaking English, alongside a large percentage of the online world means we are sheltered from the some of the issues that the search engines face to do with language.
It still surprises me when I see results that include multiple languages. I’m very jealous of the small number of people who are truly multi lingual, but ignoring that small subset of people, there is no possible reason (apart from covering all your bases) that returning results in multiple languages is a good idea.
Results must be in the right language
The quest to return the most relevant results to any query starts by trying to understand the keyphrase the search has chosen. For the search engines to return a relevant set of web pages the search engine must first work out what language the user is searching in, and secondly try to only return results that match that language. Determining what language the resulting pages should be in has a number of problems:
- The keyphrase used may not give away what language the searcher speaks. Consider a search for ‘php’, what language is that in, English, Dutch or perhaps Italian? Interestingly the italian search returns the main php website (in English) as the number 1 result. Not a great user experience if you are Italian and don’t speak English.
- You can’t rely on the tld of the search engine you are using. I could be on holiday in italy with the default engine being google.it, yet I still want my english search to return english results.
- You can’t rely on the location of the computer, just because someone is searching from france doesn’t mean they speak french.
Tips to Identify what language you are speaking
Google tends to be pretty good at working out the language of a given page, but as with most things in this industry the belt and braces approach is often best, and taking these few small steps can help ensure that your pages are having the language correctly identified.
- Include a language tag in the html. You can do this by amending the
<html>tag. This article on Dive into accessibililty will explain more. - Ensure each page only has writing in one language on it.
- If you want to have multiple languages for your website either split them into different domains (which allows you to have relevant tlds), or group the language by folder, eg apple.com/fr. Watch this space for more detailed discussion about whether you want a single domain or multiple domains to hold the different languages you have.
If you have enjoyed this post you can subscribe to the rss feed to read more about how you can monitor and protect your brand online


Comment link
Steven Tolliver on Fri (22 Feb) @ 2:57 pm
Great post! As an (American)English speaker working in Spain I’ve often found weird results for online search depending if I’m using google.com or google.es. In general google.com covers a lot of Spanish sites, but for a more complete picture, the .es site is better. Then there’s Google Maps. GMaps.com lists one sole location for my company in San Francisco (a friend’s company there has a reference to us on his site). To find us where we really are, you’ve got to search on GMaps.es. A rare anomaly, to be sure, as most map use for finding companies is local only, but it points to need to check to make sure your search engine listings are going where they should be.
Comment link
Duncan Morris on Fri (22 Feb) @ 3:03 pm
Hi Steven.
We have a similar problem in the UK without the language differences. The results for a given search widely differ depending on whether you search in google.com or google.co.uk. I can only imagine it being worse in your situtation.
It doesn’t help now, but I’m sure this is something Google is constantly working on, so I would expect the Google Maps issue to resolve itself in the future. It does highlight the issues though.