A fun mini-project: turns out despite 10 years and ~500m+ pages served a month, many from non-english speaking teachers and students, the current code.org website in 2024 has been constructed with an in-house CMS system that has unintentionally been SEO-hostile to google-bots.
So, as a first step to unveiling our amazing work delivering computer science, computational thinking, and artifical intelligence education to millions of kids in over 100 countries all over the world, today i hacked together a new microsite that surfaces details of our programs and courses in Spanish, Hindi, and Korean. This will be followed with as many as 30 additional languages, as I study the webcrawling and search traffic patterns, to best connect these searches back to our curriculum and “app”, the Code.org LMS (courses, teacher administration, professional development, and actual classroom activities for kids).
The code.org materials in Spanish are comprehensive and delightful and edited by yours truly with help from Karina my colleague, as we are both Chilenos!
This has been super fun project as I can see immediate impact with traffic in multi-lingual queries, reaching the site and much of it converting to sign-ups and create-new-account buttons. Kudos to Webflow, such an amazing website building tool, has been a pleasure to use.
Over the period of the project the search volume was pretty good and in particular the average search position was stellar (~5). However despite tagging these pages with the language code and native content in Thai, Italian, Tamil, Hindi, etc. — the primary domain of code-org continued to receive significantly more traffic even though it was serving english page results to Thai/Tamil/etc. queries. Here is the volume of search traffic to the microsite:
And by comparison, look at just the search traffic from one country Thailand, in that same period:
So quite an interesting experiment. Having a native Thai landing page, made a dent (580k impressions v 3.13 to the primary domain english page), but the click through rate was 1.6% (to the Thai page) vrs. 43% (to the english page on the primary domain).
The take-away for me is that (a) more time was needed to establish authority of the subdomain, (b) using a subdomain a huge disadvantage especially on a 10 year old, trusted main domain.
The good news is today we are ready to fold this traffic back into our primary domain and have staged new marketing website pages using a new CMS (Contentful), and that system is connected to an awesome new TMS (Translation Management System) called Localizejs. With this new TMS we are going to radically reduce our cost and time to market, and most importantly, extend coverage of translation into 22 languages for more of our student curricula and thus reach more students and teachers.
Here’s the new student marketing landing page for code.org in Thai, on the main domain. Now Thai queries should rank for our https://code.org/th/global.
For 2026 we are starting to focus on India as a very strategic market for us to expand computational thinking, computer science, and AI literacy education of K-12 students. Offering these materials in Hindi, Kannada, Tamil, Marathi, and Telugu will now be possible thanks to our amazing new AI translation workflows using Localizejs.
And here’s a general overview of all of Code.org’s programs in India.
And here is a complete list of the initial 22 languages we are supporting with this new translation platform:
Forest Key's Blog 