How We Developed A Social Media App For A US Political Community And Scaled It To 4 million Users
At APPiLY Technologies, we developed a social media app for a client based in the United States. Our client wanted to create a brand new social media platform focused primarily on politics and politics discourse. We started development work in late 2019 and launched the first version after six months of development in early 2020.
It quickly gained traction and reached peak user engagement during the 2020 U.S. Presidential elections. It garnered over 4 million registered users and received coverage in the national TV and print media.
Why Develop Another Social Media Platform?
Unbiased Platform Free from Influence
Most of the early adopters of our social media platform were right-leaning republican party supporters. Among them, there was a growing disillusionment with mainstream social media platforms. Many believed these platforms engaged in practices such as shadow banning, suppression and blacklisting to downplay opposing view points and influence election outcomes.
This platform was developed as an alternative - a safe and equitable space free from censorship and external influence.
Data Privacy
Mainstream social media platforms have often faced scrutiny over their handling of personal data. Reports of user data being monetized without clear consent, sold to third parties, or exploited for targeted advertising have raised serious privacy concerns.
In contrast, our social media platform is designed with data privacy as a core principle, ensuring that user information is safeguarded and used transparently and never used for monetisation.
Specialized Features Aligned with Community Goals
Mainstream social media platforms are intentionally designed to serve all sections of the society. They come with a wide range of features catering to every subgroup and interest. However, this broad approach dilutes the focus on specific topics or communities.
In contrast, our social media app had a singular purpose - fostering meaningful political discourse. We developed it with several specialised features (such as a public forum for political discussion, a newsroom delivering the latest political news, a video channel for political podcasts) all seamlessly aligned to creates a cohesive and focussed experience.
Customised Revenue Model
Mainstream social media platforms predominantly rely on advertising as their primary revenue source. To maximize profits, they often employ aggressive ad strategies that can disrupt user experience and dilute the platform's focus and cohesiveness.
In contrast, our platform adopts a customized revenue model that prioritizes transparency and user satisfaction. Instead of intrusive advertisements, revenue is generated through voluntary donations and subscriptions to premium content, keeping the platform clean and safeguarding user interests.
Core Features Developed
Public Forum
The main page of the app (also called as Public Forum) shows a curated list of posts from across the user community. Posts are sorted based on their popularity (likes, bookmarks) and engagement (comments, shares and replies) in real time and presented to the users.
Private Forum
This additional personalised feed shows posts only from the users friends and followers, thus containing only posts aligned with their interests. Posts can be sorted based on popularity and trends, or displayed in chronological fashion, as per the users preference.
Besides the posts, the forums contains user recommendations for following and announcements from the administrator.
Public Groups
Admin and users can create groups, which people can browse and find and subscribe to. Each group has its own group timeline with posts from within its group members. Group admin - through settings - can add and remove members, block and delete posts and enable post moderation.
State/City Groups
State and City Groups are groups based on geographical locations - for discussions around local issues, organising community events and mobilising support for political rallies. These local community groups helped the platform develop a strong community which is active both online and on the ground.
Messaging
People can engage in one-to-one private conversations with the messaging feature. It is tightly coupled with the rest of the social media - a user can click on the profile picture of users in the public forum posts, replies, comments, likes, shares and video sections and initiate a conversation.
To prevent spam and unsolicited messages, the messaging inbox is segregated into two sections - one for conversations from friends and another for the rest.
News Feed
The news feed is a standout feature where news stories from multiple sources are collected, categorised and presented to the user in real time. News items are organised into categories such as politics, sports, business, finance and health. Users can subscribe to topics of interest and get real time updates.
The news items were sources not only from mainstream media outlets but also from independent freelance journalists and small local news outlets. This gave independent voices and content creators a platform to broaden their audience and share underrepresented news stories.
Video Library
The video library hosted all the videos shared in the platform. Users can sort the videos by chronological order or by popularity and search using queries and hashtags. Video library was tightly integrated with the rest of the app, allowing users to cross post videos into forums, groups and messages and vice versa.
Video Channels
A feature similar to YouTube Channels, allows content creators to create their own dedicated channels within the app and share videos, stream live events and podcasts. This feature allowed political activists and podcasters to create their own independent content, share and build an audience.
Engagement Features
We included several tools and features to encourage greater participation and interaction among the community members.
Likes, Reactions and Bookmarks
Users can like posts, react with emojis and save the posts for future reference.
Share and Comment
Users can share posts of others to their followers and show support. Users can reply and comment on posts and participate in a thread of discussions on a topic.
Friends and Followers
Users can make friend requests to people they know or want to have a closer relationship. They can also follow people of interest and get updated when they post.
Privacy Features
Privacy and security features were integrated to the app to make people feel safe and secure in the community.
Report Posts
Users can repost posts considered to be inappropriate for the community. An alert is sent to the administrator who can review the post and take action - block the post or delete it.
Report Users
Users can report spammers and repeat offenders to the admin as well. The admin can place the account under suspension or deactivate it altogether.
Block Posts
Users can also block specific posts from appearing in their timeline, giving more flexibility to handle averse content.
Block Users
As a last resort, a User can also block other users and completely block all interactions with them. Posts authored by them will not appear in the timeline, they can’t comment or share the users post and they can’t message privately, once blocked.
Project Execution
The project was executed using Agile development methodologies with a cadence of 2-week sprints. The team worked closely with the management and design teams in the US, to define the scope and requirements for each feature and deliverables in each sprint schedule.
Technology
Frontend
The Mobile App for Android and iOS were developed using Native App Development Frameworks and Languages - Swift for iOS and Kotlin for Android. The web app was developed using Angular.
Backend
The backend server logic was developed using javascript frameworks NodeJS and ExpressJS. The backend provided the API (Application Programming Interface) which was consumed by the frontend clients.
Database
We used AWS DynamoDB to store data related to posts and content. User data and user connections were managed using PostgreSQL.
Cloud
The application was launched on an AWS Cloud. We used several AWS Cloud Services to build and scale the infrastructure.
Amazon EC2, AWS AutoScaling and Elastic Load Balancing - to run and scale the applications servers
AWS S3 and Amazon CloudFront - to store the media files and to deliver them.
AWS RDS - to manage the SQL database.
AWS Lamda - to run short duration service tasks such as content moderation.
Amazon SQS - for communication among the micro services.
Elastic Search - to provide content search capabilities.
Scaling the Platform
By Dec 2020, we had over 4 million uses on board, with daily average posts of more then 100K. To handle such a large volume, we had to scale our infrastructure, database and the backend software quickly. We employed several techniques to scale the platform and optimise costs at the same time.
Auto Scaling and Load Balancing:
Our users were spread across the United States, spanning from the East Coast to the West Coast. As the day progressed, the system experienced dynamic fluctuations in load, reflecting the varying usage patterns across different time zones. To handle these spikes and troughs in traffic, we implemented autoscaling and load balancing. Autoscaling dynamically added or removed infrastructure resources based on the system's load, ensuring optimal performance and cost efficiency.
Database Scaling
Data was segregated and managed by different types of databases—SQL and NoSQL—based on its type and access patterns. For high-performance and scalability, we used AWS DynamoDB to store data related to posts and other content. User data and user connections were managed using an SQL database, which was scaled through database clusters with load balancing for read workloads.
Micro Services Architecture
To achieve performance at scale with cost efficiency, we split the backend software into multiple independent services - each capable of scaling independently. Tasks such as processing new posts, checking for profanity and obscenity in content, processing media files (such as photos and videos), and ranking and distributing posts were handled by separate services. These services communicated and exchanged data using a scaled shared queue mechanism.
4
Million
Total Registered Users
100
K
Average Daily Posts
3
Million USD
Funds Secured
6
Rank
Top App Downloads
Contact Us
Enter Your Name
Enter Your Phone Number
Enter Your Email
Briefly Describe Your Requirement