Translations and marketing segmentation - How to leverage language in your email campaigns

3 min read


Marketing segmentation for emails is difficult to get “right”.  When you add multilingual content into the mix it gets even harder.  One of the biggest pitfalls, however, is confusing language codes for marketing segmentation. In this blog post I’m going to talk about why these should be treated differently and offer some strategies for dealing with them.


So let’s start with a simple fact: languages are not countries.  We often “confuse” country and language and treat them as the same.  While there are many cases where is a “safe” assumption, there are many where it is not. In the US alone there is over 30 million Spanish speakers, its pretty clear that targeting everyone in the US “segment” with English content is going to lead poor engagement. This is the first part of the problem; split out language segments separately from geographic locations.


Next fact: Languages are standardized while segmentation is made up. In order to ease translation, languages are assigned a specific set of codes, these are often referred to as locale code.  These codes are important to translators. They help the translators to determine exactly what language you are asking for. Conversely your segmentation doesn’t have to follow the same rigor. In fact it’s fairly common to see a single geographic segment actually have multiple languages.  Often this leads well intentioned marketers to make up new codes. Going back to our example in the US where you are targeting Spanish speakers. It might make sense to have a es-US marketing segment to represent this segment, but avoid sending this code to the translator. Instead tell them the standardized language code you want (probably es-MX).


Hopefully you are convinced that marketing segment and language should be treated differently, so now lets look at a couple approaches that gets the right language to the right segment and results in higher engagement.

  1. Keep a “mapping” of language locale code to each segment. If you think about it, every email *has* to have a single language (well there are cases for bilingual content, but thats a topic for another blog post).  Because of this we can map every segment to a language. Now this approach might be a bit redundant because most of your segments might be the same. So build your mapping in the manner that keeps it simple to update and refer (ie only include non-English mappings for example).  Also include a designation in the segment name so that you know that you should refer to the mapping.

  2. Separate the content from the email segmentation to allow language to be independently associated.  In this solution we don’t *have* to make language segments (unless we want to). Instead language can be determined by the subscribers information and our email templates fetch the associated language content. I like this solution because now we can use language locale codes and not have to worry about impacting our segmentation design. However for this solution to work our email templates need to become *smart* since they need to handle fetching the right language.

At Cloudwords, our customers use many different segmentation strategies.  So we are accustom to supporting large enterprises marketing around the globe.  That’s why its critical that we don’t act as a cloud service provider, but also as a partner.  This ensures our customers are always well supported whatever approach they take.


Written by Matthew Jackowski