Speed dating 2.0 has been proven an effective way to improve user engagement and monetization by several pioneer online dating platform. We will explore how to effectively implement one using ZEGOCLOUD services.
Challenges to Online Dating Platform
To run the speed dating 2.0 business, pioneer online dating platforms have experienced and racked up a few pain points:
1) Chances are very high that the dating pair matched doesn’t like each other.
2) It takes ages to open an audio or video chat room to start a conversation.
3) There are pausing, lagging, or even disconnecting issues during a conversation.
4) For known reasons, inappropriate contents bring potential legal risks for platforms.
In this article, We will focus on the standard practices to fix these pain points.
How to make an excellent online dating platform match
There are several standard matching models in the market. You might use one of them or combine some of them to develop an innovative model.
1) Matching by attractiveness
Request the two users to upload a video clip of themselves on the online dating platform. Make sure that the video clips include their heads and faces. There ought to be an attractiveness evaluation model. The model is built into an algorithm to rate a person’s attractiveness with quantitative metrics. Appearance style and attractiveness level etc are factors to consider.
2) Matching by a shared interest
It is said that two persons don’t have to have shared interests to start a relationship. However, common interests are catalysts to help them find something in common to talk about in an ice-breaking way. It creates an opportunity for them to be fond of each other.
If the shared interests are something they can do together online, such as online mini-game and online karaoke, the system can guide them to launch a mini-game to play together or enter an online karaoke room to sing a song together. These measures can help the two persons build rapport quickly and create opportunities for the platform to provide premium services.
3) Matching by geographic location
It is a clever way to match by location in an online dating platform. The geographic location dimension often means a great deal:
a) Opportunity to upgrade to blind dating for marriage;
b) Chance to upgrade to premium products and services;
c) Shared culture and intimacy feeling.
A matching model cannot ignore this factor.
4) Matching by preference
Understanding why a person likes or dislikes another person unless the person says it aloud is highly complex. User preferences are essential data that can reflect why and what. Setup at least two dimensions for preferences, i.e., likes and dislikes. Then you can include candidates by their likes and exclude candidates by their dislikes.
How to enable entering a chat room instantly
After a matching is done, users cannot wait to discover what the surprise will be. The matched pair then joins a voice chat room for a conversation. The period from clicking a join button to hearing the other person’s voice or seeing the person’s video is a test of users’ patience which is very short. We have to make it as short as possible. Let’s use ZEGOCOULD’s client cases as examples, we manage to make the period shorter than a second, and our clients call this great feature “open-in-a-second.”
1) Optimize the Login Process
There are two concepts we have to understand first:
It is a virtual space to group the room members for privacy protection. Only room members possess permission to publish or playback streams.
You have to publish a stream in the room to allow the other room members to hear or see you. Likewise, to hear or see other room members, you have to subscribe and playback their streams. Publishing and playing back streams are called streaming.
From clicking a join button to hearing or seeing the other person, it involves steps including room login and streaming. Let’s see how to shorten the period of room login. As a common practice in the market, the room-login process will include the following steps:
- A user app submits a login request to a room manager server;
- The user app receives a triumphant return from the room manager server. The user app can obtain a destination IP from the return and publish a stream to a media server identified by the destination IP.
- The user app submits a login request to a heartbeat server that can maintain heartbeat testing even if the user app gets disconnected later.
- The user app receives a triumphant return from the heart-beat server.
The room-login process takes about 610 milliseconds on average.
ZEGOCLOUD has made optimization of the room login process
It includes the following steps:
- A user app submits a login request to a room manager server;
- The room manager server submits a login request to a heart-beat server on behalf of the user app.
- The room manager server receives a triumphant return from a heart-beat server on behalf of the user app.
- The user app receives a triumphant return from the room manager sever.
After the change, it will take about 180 milliseconds on average.
2) Optimize for parallel streaming
Once the two users accomplish room login, they can start streaming.
As a common practice in the market, the streaming process includes the following steps:
- User A requests for stream publishing and receives the IP address of a media server (Let’s call it media server A) from a scheduling center.
- User A establishes a connection with the media server A.
- User A starts publishing a stream.
- The room is updated with a newly added stream entry.
- User B is notified of the addition of the new stream entry.
- User B requests for stream subscription and receives the IP address of a media server (Let’s call it media server B) from the scheduling center.
- User B establishes a connection with media server B.
- With the help of the scheduling center for streaming back-sourcing and routing, User B starts playing back the stream published by user A.
User B will get user A’s stream only after a stream change notification is broadcast in the room.
Users cannot avoid waiting time. Some parts of User A’s actions and user B’s actions can be done in parallel.
ZEGOCLOUD has optimized the streaming process for more parallel actions.
We will emphasize the changes in the following steps:
- Once user A starts requesting stream publishing, the room is updated with a newly added stream entry, and user B is notified accordingly.
- User A will continue the steps accordingly.
- User B will start requesting stream publishing and continue the steps accordingly.
- When user B starts playing back the stream that ought to be published by user A, the stream might not be ready yet. However, user B can playback the stream once it is ready.
It will reduce the time to start streaming by about 30%, significantly improving the “open-in-a-second” demand.
3）Global Multiple Scheduling Centers
Speed dating 2.0 is designed for international social networking scenarios. We must consider how to ensure “open-in-a-second” under international circumstances.
Suppose there is only one centralized scheduling center for the data transmission network; what would happen? All scheduling requests must go to that scheduling center, which will become a traffic bottleneck and introduce unbearable latency for some countries far away from it.
ZEGOCLOUD deployed multiple scheduling centers, with exclusive scheduling centers for some countries with significant population numbers. The multiple-scheduling-center strategy ensures that scheduling time is optimized and the “open-in-a-second” goal can be achieved.
Low-latency real-time audio or video
There is concern that the quality of real-time communications cannot be guaranteed under weak network circumstances. Network infrastructures in some developing countries are inferior, and cross-border network connections are even worse. Ensuring quality user experience for real-time audio and video communications is paramount for the success of the speed dating 2.0 business model.
ZEGOCLOUD ensures the quality of real-time communications from two aspects.
1) A global data transmission network with more than 500 BGP network nodes
The relationship between data traffics, and infrastructures are like that of car traffic and road networks. Good road conditions can help cars to go smoothly and fast; the same applies to data traffic and network infrastructures. ZEGOCLOUD has deployed more than 500 BGP network nodes in 212 countries or territories. The data transmission network has laid a foundation for global real-time data transmission.
2) A network bandwidth adaptive algorithm to handle dynamic network congestion
ZEGOCLOUD has built a real-time transmission model that can adapt to network condition changes and ensure smooth and low-latency real-time communication.
In this real-time transmission model, the sender end can predict network congestion and bandwidth by the feedback from the receiver end, and then it can adjust the packet size that it encodes and transmits. As a result, the bandwidth that the sender end uses to transmit data changes dynamically and is always lower than the available bandwidth.
How to discover inappropriate content automatically and at a low cost?
In the speed dating 2.0 scenarios, users speak with each other in a one-on-one manner. Chat rooms in an online dating platform are private virtual spaces, and intimacies will build up as relationships develop. We must face a challenge to protect users’ privacy and align with local legal compliance.
ZEGOCLOUD works with top AI-powered content auditing solution providers to automatically identify inappropriate audio and video content.
We conduct automatic content auditing on the server side to ensure that content is done at a meager cost. and we have taken two measures to reduce costs.
1) Mixing streams to reduce bandwidth
Streams from users A and B are mixed to produce a single stream for automatic content auditing at runtime. This measure can reduce bandwidth costs and make content auditing easier.
2) Mixing streams to produce recordings
The fixed single stream can be saved as a recording file. When there is a need for review by compliance staff or regulation officers, they can see the videos from both users A and B in one single file.
If you want to motivate your platform users to pay for premium services, speed dating 2.0 is the right choice for you.
Please contact us to discuss with a ZEGOCLOUD expert for advice on implementing a speed dating 2.0 business.
Talk to Expert
Learn more about our solutions and get your question answered.
Take your apps to the next level with our video APIs
- 10,000 minutes for free
- 4,000+ corporate clients
- 3 Billion daily call minutes