Facilitating Tandem Sessions— UX Case Study

How I used UX Research to identify and come up with solutions for Tandem to deliver on its promise of helping people learn/master a foreign language and reduce the platform’s churn rate.

Tandem is a language learning app where millions of people teach each other.

“As many languages you know, as many times you are a human being”
Kolik jazyků znáš, tolikrát jsi člověkem. (Tomáš Garrigue Masaryk)

→ Leia este artigo em português

About

This case study was developed in the period between December 2020 and January 2021 and is structured in the following topics:

• Challenge
• First-hand knowledge
• Desk research
• Planning the interview
• Selecting candidates
• Findings
• Persona & User Journey Map
• Problems
• Ideas and functionalities
• Bonus solution
• User interface

Challange

Understand why people stop using the app and come up with solutions to reduce the platform’s churn rate.

First hand knowledge

As a former user of the platform, I listed pain points that I experienced myself. They are:

1. Language settings (native, speak, learn) would affect the discovery settings in a non-intuitive way, making me see people who don’t want to learn my language or speak the language I want to learn.
2. Uncommiteted users who aren’t willing to learn or teach a language, making it hard to find suitable partners and perceive the platform as a viable option to learn a language.
3. Chat system with bugs.
4. Translate feature adds no value to the Pro account if the user knows about Google’s Tap to Translate.
5. Lack of clarity about the version of the language people speak or learn. A person who learns European Portuguese might not want to practice with a Brazilian due to the difference between Brazil and Portugal and a lot of time can be wasted just to figure out which version the other person speaks or learns.

Desk research

Before anything else, I took a peek on the internet to see what people were saying about Tandem — the app and the learning method.

For the sake of brevity, I’m listing below four links where I found interesting inputs about the matter:

1. Tandem reviews on Google Play
2. 5 Reasons Why Tandem Partners Don’t Work
3. I Don’t Like Tandem (And Similar Apps)
4. TANDEM REVIEW: A TINDER-ESQUE LANGUAGE EXCHANGE APP?

Planning The Interview

A semi-structured interview for a qualitative research will be ran via Google Meet with inactive users.

Goal

To find out why some users go inactive on Tandem.

Objectives

1. Learn the user’s history with the language(s) so far and emotions involved;
2. Find out about their overall experience with Tandem.

Hypothesis

• Tandem doesn’t deliver on its promise of helping people master a language as it is just one tool of many.
• It’s difficult to find people who are really interested in doing tandem to allow evolution in the target language.
• As an app that offers nothing but access to foreigners and a messenger with quite a few bugs, Tandem gives one no tools or reasons to stay once they finally find compatible people who they can practice or build a relationship with via other messengers like Whatsapp or Telegram.

Selecting candidates

I’ll send people who I kept in touch after quitting the app a link to a Google form that consists of a conditional question: How often do you use Tandem?

The following screen(s) will depend on their answers. Here’s how it’ll work:

How they’ll be classified

Users who are not currently using the app, but “might use it again in the future” will be considered Inactive users because they serve the platform no purpose. They don’t:

1. Generate revenue through subscriptions or ads;
2. Help other users, who might end up seeing no value in the app due to lack of collaboration, the essence of tandem partnership, and leave.

Frequent and sporadic users will both be considered Active users.

The following diagram shows how the preliminary results will be treated.

Preliminary results

None of the users I kept in touch with who were pretty active by June 2020 is a frequent user now (Jan 2021). Some of them only open it in the rare occasion of receiving a message, in order to not let the person hanging, not using it as a learning tool anymore, and others have uninstalled it and kept it touch in other platforms.

The numbers are:

• Frequent users: 0 of 10
• Sporadic users: 4 of 10
• Not current using: 4 of 10
• Won’t use it again: 2 of 10

Four is still a good number for a qualitative research and so I proceeded with it. Let’s explore our findings now!

Findings

The process confirmed there are people who aren’t really there to learn a language, but to connect. These tend to move to other messengers afterwards.

I had some insights from people whose goal was to learn. For example:

1. Social anxious people see in Tandem a chance of overcoming their fear of connecting with strangers, but not knowing how to start a conversation or how to keep them going, specially in another language, can be a burden.
2. Two non-natives could benefit as “collaborative classmates” from using the whole session to focus on a common target language.
3. The platform’s structure stimulates synchronous interactions, which can be a problem for tandem partners who are in very different time zones.

For all of them, though, one common aspect is present: the struggle to organize and maintain sessions and partnerships. It seems like the platform should come up with solutions to facilitate tandem sessions in order to help users accomplish their goal (or milestone) of learning a language.

Persona & User Journey Map

Problems

On language learning as a whole:

1. Learning a new language online can be like looking for a needle in a haystack — people use way too many apps/resources with no clear roadmap/method.

On Tandem:

1. Lack of structure: Students don’t know where to start, what to study and how to help each other to make the most out of it.
2. Lack of understanding of what tandem is as a method.
3. A1 and A2 don’t seem like productive levels for tandem.
4. Time difference between partners.
5. Hard to find users who are really committed. A lot of them even flirt and make it Tinder-ish.
6. Difficulty to make partnerships last because users stop using the platform frequently and one has to find new partners quite often, which can be exhausting in general and distressing for people who are socially anxious and find it difficult to build trust and feel comfortable with others.

Ideas and Functionalities

All ideas and functionalities aim to tackle two main issues:

• Increase trust within the community.
• Help users start and maintain productive tandem partnerships.

They are:

1. Identity verification
2. Tandem request & Set limit of partners
3. List of interests for affinity sorting and conversation topics
4. Formal-Informal Language switching
5. Tap to Translate & Turn into Flashcard
6. Database of texts to work in tandem
7. Party Games
8. Pen Pal

Identity verification

Verified User label to gain trust within the community, making sure users can’t cheat the system by deleting all their photos and then changing to someone else’s photos once they get the label.

Tandem request & Set limit of partners

Prevent users from being overwhelmed by messages. While the profile would stay public, the message system could allow more privacy. If the user doesn’t want to set a limit, then messages would keep working the way they do now.

List of interests

• Affinity feature: Allow users to inform their top ten interests to allow sorting results by common subjects and increase chances of long lasting connections where users context is brought to the learning process, which is what Paulo Freire — patron of Brazilian education — defended to help illiterate people to learn how to read.

Users shouldn’t be able to see interests that don’t match in order to prevent editing their own interests to be someone’s “perfect partner” and distort it.

• Conversation topics: When clicking a specific common interest, suggested sentences would be displayed to help start the conversation or keep them going. From that alone, a lot of grammar related questions might come up and the participants would go with the flow.

Formal & Informal

Allow users to switch recommended sentences from formal to informal language and vice-versa if that’s important in their target languages.

Tap to Translate & Turn into Flashcard

1. Allow users to turn any word on the device into a flashcard with the context the word was used in by copying the word (smartphone) or browser extension (desktop).
2. Gamify this feature by challenging the user to get it right a few times a day via push notifications. About five cards every round so they can find time during a busy day. The more the user gets it right, the less it’ll appear to the user. Again, it’s important to stress how much they advanced instead of bringing them the feeling of loss.

A benefit for the business is that users would be concentrating all their learning in the platform, increasing fidelity.

Database of texts

A database of texts of all levels about all sorts of subjects (Sports, culture, history, geography etc.) to facilitate sessions. Again, the idea is to bring user’s context to help them navigate the target language more easily.

Party Games

Games to help users get to know each other, loosen up, feel comfortable, have fun and build both trust and a relationship while learning a language, like:

• Trivia games
• Forehead detective
• Dark games
• Scrabble

Pen Pal

Time difference can make it difficult for partners in different time zones to find room in their schedules to practice live, which can lead to consecutive postponements and ultimately to the connection being broken.

Pen Pal would allow them to simulate the old days letter exchanging habits with all the great features of correction, turn to flashcard, read for the partner to develop their listening skills and more.

Bonus solution

This is the solution I found too complex and expensive to implement in short or medium term — also it could have nothing to do with the business objectives — but to me it sounds like the only one that would definitely help users find committed people so I’d like to mention it anyway.

Maybe it could work as a partnership with another platform. Let’s see what it is about. Who knows what team work could make of it, right?

Course Platform (The Udemy of language learning)

A self-regulated course platform with videos, podcasts, texts etc. produced by independent instructors of all backgrounds and styles who’d be reviewed by students.

This way, Tandem would allow a variety of methodologies so students can find the one that suits them better whilst focusing on developing great features to make it an efficient tool for language learning and turn it into “all you need to learn a language, from courses that would give you a clear roadmap to natives with whom you can speak and master a language with!”.

Students would have the option to

• Pay per Course
• Subscribe to Tandem Pro Plus
• Search for people who are enrolled (and actually watching the lessons)

Tandem could populate their YouTube channel with crash courses, let’s say Level A0, starting with the most learned languages, to promote the business via content marketing and bring users to buy courses and become someone’s tandem partner.

User Interface

I’ve just started working on a couple of tweaks to make it look and feel more modern and integrate the new functionalities I came up with during this process.

I’m currently experimenting the use of symbols to communicate certain areas in the app, having in mind the consistency of the UI for it not to break depending on the language since some have more lengthy words than others or a right to left system for instance, but only a usability test would tell us whether or not it’s the way to go.

Check out a brief demonstration of it so far and I’ll be updating this session in the future.