Protect Users from Direct Message Harassment Under the Anonymous Context

Company
Fuye Tech.
Time
2 weeks , Mar 2020
Current State
Launched
Team
Product Designer (Zane Hu), Product Designer (Sylvia Liu), Product Manager (Samuel Chen), Engineer (Steve Fu), Engineer (Liqi Chen)

􀉚

Overview

Product Brief

􀫕

􀫕

Zhitiao (Whisper)

👤

Anonymous

Anonymous

🏫

Campus-centric

Campus-centric

💬

Social Media

Social Media

PROBLEM

􀇾

􀇾

🗯️

🗯️

The anonymous nature of social media has led to an increase in direct message harassment complaints.

Design approach

􀬁

􀬁

🔄

🔄

Transfer the power to initiate conversations from the sender to the recipient.

SOLUTIONS

􀋭

􀋭

🗣️

Sender

Restricted Content for First Message

Senders are limited to crafting their initial 'Say Hi' message from a certain set of phrases. The chat can only continue when the recipient chooses to reply.

Complete Flow

Key Screens

Craft Content
(Choose Real ID / Anonymous)

Wait for Reply


Conversation Blocked


👤

RECIPIENT

Power to Approve Conversation Continuation

Recipients can choose whether to reply or decline the messages. Only when they decide to reply will the conversation proceed.

Complete Flow

Key Screens

Choose Decline or Reply

Decline

Reply

Imapct

􀐾

􀐾

-42%

The rate of reporting DM harassment

🥰

Received lots of positive feedback from users

􀉚

User research

Even with Reporting and Blocking features in place, we still got complaints about DM harassment. So why is that?

􀌬

􀌬

204

Online Survey

I conducted an online survey within the app and gathered quantitative data from 204 responses, which confirmed the severity of the problem.

8

User Interview

I also screened 8 users who had experienced DM harassment and conducted in-depth interviews with them to dig out their real thoughts.

key insight:

􀛭

􀛭

💔

What is needed is prevention, rather than remedy after the hurt.

"Although I can report and block the person afterward, I've already been hurt by those words."

HOW MIGHT WE…

􀇯

􀇯

☔️

Reduce DM harassment by providing a set of effective prevention mechanisms

􀉚

Explore Mechanisms

So what other platforms do to prevent DM harassment?

􀁜

􀁜

I analyzed what other platforms' prevention mechanism, including RED, TikTok, Weibo and more.

There are 2 common mechanisms:

🔒

Limited Message Amount

The sender can only send one message unless the recipient replies.

Customized DM Preferences

Users can customize their DM rules such as "only receive DM from people I follow".

Better experience makes loyal customers

But, they DON'T work for our platform because:

􀯨

􀯨

🔒

For Limited Message Amount:

Users have to send lengthy messages to clarify intentions, which can be cumbersome on mobile devices. Plus, harassment could still happen for the first message.

For Customized DM Preferences:

As an anonymous app, maintaining anonymity is crucial. It makes distinguishing between users impossible.

Though, inspired by them, we started exploring our own

􀯧

􀯧

Dump Ideas & Affinity Mapping

We ran a 'Dump Ideas' session, after which we sorted ideas and discovered they could be categorized into 2 types.

Finally, 2 mechanisms were chosen and refined:

🗣️

Sender

Restricted Content for First Message

Senders are limited to crafting their initial 'Say Hi' message from a certain set of phrases. The chat can only continue when the recipient chooses to reply.

👤

RECIPIENT

Power to Approve Conversation Continuation

Recipients can choose whether to reply or decline the messages. Only when they decide to reply will the conversation proceed.

􀍕

Design development - Sender

🗣️

Sender

Restricted Content for First Message

User Flow & Corresponding Pages

􀙫

􀙫

Key Screen: Craft Content

Design Decisions

􀛭

􀛭

DESIGN DECISION #1

Use Half Sheet

WHY: Since it's a light interaction, half sheet helps to keep the user experience smooth and reduce any sense of interruption.

Open in Full Page

open in Half Sheet

DESIGN DECISION #2

Switch to a First-person Tone

WHY: Given the limitations of message content, using a first-person tone is a strategic choice to create a more immersive experience.

Second-person Tone

First-person Tone

DESIGN DECISION #3

Update the Component for Switching Anonymous Mode

Original

WHY: The component is visually boring and makes the decision of choosing it unimportant.

Toggle Switch

WHY: It's intuitive, but not highlighted enough. Also, it's a bit weird visually.

Radio Button

WHY: It makes this important option less obvious and feels like a form to fill.

New Component: Avatar + Name

WHY: Adding both user's & anonymous collective avatars and names highlights it and enhances the immersive experience.

old

New Component: Avatar + Name

Usability Test

􀟲

􀟲

👤

Problem 1:

25% of them asked if they could choose to not display any personal information while in anonymous mode.

😵‍💫

Problem 2:

The visual difference when switching between Real ID and Anonymous mode is not obvious enough.

TWEAK #1

Make sharing of personal information optional

old

new

TWEAK #2

Color Changes when Switching

old

Color Changes when Switching

All Screens (Sender)

􁣴

􁣴

􀍙

Design development - Recipient

👤

RECIPIENT

Power to Approve Conversation Continuation

Design Decisions

􀛭

􀛭

DESIGN DECISION #1

Moving the task of reading the message after the recipient decides whether to reply to the sender or not.

Initial User Flow

New User Flow

DESIGN DECISION #2

Add the original post for clearer context

WHY: It was difficult to understand the context of the DM, as users couldn't frame it using restricted content.

old

new

All Screens (Recipient)

􁣴

􁣴

􀉚

Final Outcome

Flow of the Sender
(Restricted Content for First Message)

􀍕

􀍕

Choose to Say Hi

Craft Content & Send

Wait for Reply

Case #1: Conversation Blocked

Case #2: Continue Conversation

Flow of the recipient
(Power to Approve Conversation Continuation)

􀍙

􀍙

Choose Decline or Reply

Case #1: Decline

Case #2: Reply

􀉚

Takeaway

💡

Customization Over Convention

While established design patterns are helpful, they’re not definitive. It's crucial to design features that are not only familiar to users but also uniquely attuned to our product’s specific context and user needs.

🧭

🧭

Validate Uncertainty through Soft Launch

Creating an innovative solution was exciting but filled with self-doubt. By soft launching, I was able to quickly validate it and uncover potential issues. It ensured that the final design not only solved the problem but also is accepted by users.