Combining social interaction with language learning

Designing a Grammar Cheatsheet

Learning a language could be fun and painful at the same time. To be able to enjoy communicating with others in a new language, sufficient practice is an essential part. In traditional SLA(Second Language Acquisition) approach, the tutor can try different variations of practice activity to make it more fun and personalized, but for most E-learning products (including Babbel), the practice activities are usually repetitive spelling drills. 

The goal of this research is exploring the connection between social interaction and user engagement, by redesigning the current practice feature in Babbel. 

PROJECT
Master thesis research with supervisor from University of Twente and Babbel
MY CONTRIBUTION
Researcher and designer for the independent project
T_cover
Final prototypes for A/B testing with the variable of chatting to real user or the software
a screenshot of Grammar Book feature on production
a screenshot of Grammar Book feature on production

The research process

The research process is inspired by Double Diamond Design Process model and adapted to the requirement of a thesis research. The primarily used design methods are taken from participatory design.

The project started from the trigger (the design problem): low completion rate of practice activity in Babbel product, especially within the age group 18-24. In the Understand phase, multiple direction and ideas have been explored by context research and literature study. 

Then by doing Generative Interview, the design vision has been defined: adding real-life communication in the practice activity. During Explore and Create phase, the design solution has been iterated twice by fast prototyping and user testing.

T_process
the research process confirmed in the end

Why are practice activities boring to users?

To narrow down the problem, the first step is to expand it and measure it from all different perspectives. 

T_fieldresearch
An overview of related theory and background context research

The context research gives me a broad understanding of the topic to list down the questions I would like to ask for Babbel users. Following the participatory design process, Generative Interview is an efficient method to defining the design problem in the early stage. 

The questions are divided into 4 parts with a toolbox and designed canvas.11 participants took the 30 mins interview separately (2 pilot tests). The analyzed results indicate that: Lack of real-life conversation is the key reason of lacking motivation in practice activities in Babbel product.

T_generative interview
Canvas and summarized results page from participants

Adding social interaction into the practices

One of the frequently mentioned pain point from the generative interview participants was 'I would like to practice something that I could use in real life,' in the first fast paper prototype, I focus on guiding users making conversations while practicing. 

A 30 minutes user testing has been designed based on the feature of the paper prototype.  The testing session includes 3 parts with Invison prototype; each part has one or two focused testing feature, in total 11 participants joined the interview. 

By discussing with our learning experts and developers, we came up with 4 template that works for our current content scope, also keeps the scalability for potential new content style in future iterations. (see picture in the end)

Paper prototype and the summarized user testing findings

Social interaction = higher user engagement?

Since the new redesign of Babbel practice feature changed from many perspectives such as the content, practice format, and the general interface of design language. To study the influence of social interaction separately, an A/B test with limited variable has been set up.

There are different levels of social interaction in the redesign, to keep a similar user experience, the experiment will focus on comparing the influence of direct social interaction.

T_socialinteraction
Spectrum of social interaction types listed from 'indirect' to 'direct'
T_abtesting
2 prepared prototype for comparing the influence of social interaction in A/B test.

Because of the time and resources limitation of the project, the prototype was done by using Invison without allowing the participants actually 'use' the product.

The A/B test has collected 44 survey results in total; both prototypes received positive feedback for 'helping me feel more confident to order a coffee in real life' (the content of the dialog), but there is no significant difference comparing the results. 

So at the end of the research project, we can't say that social interaction can improve the engagement of the practice activity, but the research still shows an exciting direction of adding more conversational elements into E-language learning platform.

Learnings & Shout-out

It's very fortunate for me to have the chance doing a 10-month independent research project with guidance from both university professor and mentor from the company.  Applying design theories and methods from classes into practical problems with time and resources limitation is challenge and fun.

This research wouldn't be finished without the help I got from all the Babbelonies who supported my project besides their busy work, and also I want to give my special thanks to my professor Mariet who patiently guide me through the whole research process.