Languages beyond English offer tantalizing new growth markets. I was the sole product designer on the Internationalization Team at Quora from 2017-19. Part of the job was getting creative with problem solving because it was a small team with limited resources.. This case study breaks down how small changes in user interface can drastically fuel growth. Hope you get inspired!
Best, Shakti Mb
Quora Alumni, Current Growth Designer at DoorDash
The Internationalization Team
Quora is a Q&A platform on a mission to democratize knowledge. However, language can be a real barrier to accessing knowledge. Nearly 63% of the Web’s content is in English, but only 17% of the world speaks it.
My team’s purpose was to facilitate knowledge-sharing in languages beyond English. The Internationalization Team also included two engineers, two data scientists, and two product managers.
Team results during my tenure:
- Scaled the product from 4 to 27 languages.
- 3x our non-English user base
- 3x engagement rate (time spent on the platform)
This incredible growth was directly attributable to how quickly we could design and ship experiments.
The Multilingual Opportunity
Our data scientists discovered that multilingual users were a key growth lever. Users who consumed content in more than one language accounted for 80% of content creation in the non-English languages. If we acquired more multilingual users, we could further increase content created in non-English languages.
One of our critical touchpoints for multilingual user growth was called the “Language Prompt” that asks users if they want to add another language to their Quora experience (see screenshot below). The prompt was shown to any user who had a high likelihood of speaking another language based on location, browser and device language. Our data scientists identified that even a slight increase in the prompt's conversion rate could drastically alter our growth trajectory in our international user base size.
The Experiment: Redesign the language prompt
- Business goal: Increase conversion rate of prompt to switch languages
- Team: 1 data scientist + 1 designer/dev (me!)
- Approach: A/B test
- Timeline: 2 weeks
Given I had a timeline of two weeks to design and implement an experiment, I had to get scrappy with identifying why the prompt had such a low conversion rate. Having to get creative with identifying user concerns, on Twitter we observed lots of tweets from users complaining about the language prompt (see screenshot for example of such tweets). In addition, our community managers were getting emails from people complaining about the persistence of the prompt.
The User Problem:
Using public tweets and feedback emails, I could distill the user problem into two core buckets:
- “Why are you assuming I speak German?” Users were offended by our assumptions that they speak certain languages. Languages are a sensitive topic in certain territories, thereby making our assumptions even more offensive
- “If I say yes, will German replace English?”Users were concerned that another language will replace English as their default language on Quora
The Redesign Goals:
- Make it clear that nothing will change: English will remain their primary language, and German will just be “added” to their Quora experience
- Be sensitive about how we framed our assumption about users speaking another language
What We Tested:
The Design Changes:
- Don’t make assumptions:
- I framed the prompt as a question (“Do you speak German?”). This way, it doesn’t feel like we’re making an assumption about the user, we’re simply just asking them.
- Changing the copy from “Because your profile indicates you logged in from a German-speaking country" to → "Because your profile indicates you may have been active from a region where some people speak German”. Even though the new string was more wordy, it was worth the tradeoff because it made fewer assumptions.
- Add reassurance: Reminding users that nothing in their experience will change, "English will remain your primary language".
- Personalization: Based on previous experiments, we found that when users see their own image and first name on a UI, it feels more personalized to them.
- Utilize the new design component: Our platform team was rolling out a new component for a button, so I replaced the old button with the new style button style.
- One language only: Based on a previous user research study, we found that users find multiple languages in one UI jarring. To prevent that experience, I kept all the copy to just one language in order to respect their current context.
The Language Prompt experiments led to a 258% increase in conversion to switch languages, and a ~120% increase in time spent on the platform.
Internationalization can drive growth, however, languages are also nuanced and sensitive. The dynamics of how languages are spoken around the world is complex, and software needs to reflect that.
In this case study, I shared about growth teams focused on internationalization, collaboration between design and data science and how small changes in user interface can drastically fuel growth by way of experiment design.
About the Author
Shakti Mb is a community facilitator and growth designer at DoorDash in NYC. She is an alumni of Automattic, Quora and Spotify.