How we launched voice messaging in corporate messenger
Timofey TrubenevI was responsible for voice messaging as a product designer, managing all stages from concept to delivery.
About the Product
Frisbee is a corporate messenger designed for team collaboration within large organizations. Currently, several major corporations use the service, including our company, which houses a dozen product teams. One of the primary objectives of the product is to enhance work communication efficiency.
Why We Implemented Voice Messaging
Business Goal
To fulfill a contractual obligation on the roadmap within 2 months and meet the demand from potential clients piloting the product.
Mission
To help users quickly share information using voice.
Audience
All application users utilize chats, so we anticipated that voice messages would become part of chats, thus targeting all application users. This encompasses individuals aged 18 to 55 from various fields, ranging from factory workers to IT corporation owners.
Success Criteria
At the very least, not to decrease overall user satisfaction with the product. Ideally, to increase it. We will monitor the Adoption Rate of voice messaging.
Discovery
At this stage, I explored how to approach the task by talking to users about their experiences and use cases for voice messages. I conducted desk research, analyzed competitors, and studied best practices.
Competitor Analysis
Firstly, I examined how competitors implement voice messaging, both in public and corporate messengers.
I noted several key aspects:
1. The mechanics of voice recording: tap or hold button.
2. Differences in UI bubbles.
3. Accelerated playback either within the bubble or from a panel above.
Interviews and Desk Research
My Goals for User Interviews Were as Follows:
- Understand when users prefer to record voice messages instead of typing text messages.
- Identify the most common actions when recording and listening to voice messages.
People often mentioned recording voice messages while on the go, when it’s inconvenient to type because their hands are busy or they are wearing gloves. There were several mentions of using voice messages to convey large amounts of information, such as instructions or explanations for a colleague, which are difficult to write quickly. In amusing situations, users mentioned that it's easier to convey emotional tone, like yelling at a colleague, which is hard to express in text :) Additionally, users frequently noted the use of speed-up playback and voice-to-text transcription when listening is inconvenient.
Discovery Result
After the discovery phase, I had many ideas. I outlined the best ideas with rough designs to facilitate prioritization and evaluation with the team.
Ideas and Hypotheses
All ideas and hypotheses were aimed at making the interaction with voice messages clear, pleasant, and covering the most important jobs that arise when dealing with voice messages.
1. Voice message recording
Initially, I started thinking about the entry point for recording a voice message and the recording process. I understood that our users constantly use public messengers in parallel. Therefore, the first ideas were to create a similar voice recording mechanic to avoid retraining users and make the recording familiar from public messengers. This led to the following mechanics: starting recording by holding down a button, swiping up to continue recording without holding, and swiping left to cancel recording.
At this point, I could have stopped and shown the concept to developers and stakeholders. But I wanted to rethink the concept and try to differentiate from familiar solutions while maintaining a familiar UX and making the solution more emotional. After numerous iterations, I came to the following solution:
I immediately started testing with users to understand two points:
- Do users understand how to record and cancel the recording of the voice message?
- Which version do they personally prefer based on their own feelings?
As a result, the more emotional option not only caused no difficulties for users but also was preferred by the majority of users.
I showed the solution to the developers to understand how difficult it would be to implement the second option. And wow! It turned out to be much simpler, as it had fewer interactive mechanics that could complicate development. The developers were also enthusiastic about the idea.
Then I continued to work on other important jobs and formulate hypotheses.
2. Speeding up voice message playback
3. Voice message drafts
4. Player
5. Voice to text
6. Actions on voice messages
Prioritization and Scoping
At this stage, it was crucial for us as a team to determine the version that would go into production and meet the goals of both the business and users.
I scheduled a meeting with the product manager and engineers to evaluate all hypotheses based on two parameters: technical implementation complexity and value for the business and users. We divided one by the other to obtain a high-level prioritization coefficient.
First Version
As a result of scoping, we included 4 features in the first version:
- Voice message recording
- Voice message playback
- Speeding up voice message playback
- All current possible actions on messages for voice messages.
The solution was developed for 4 platforms: iOS, Android, Web, and Desktop.
Hi-Fi Design of the First Version
Detailed designs were created for iOS, Android, Web, and Desktop platforms.
Various details and bubble states were accounted for.
UX Tests
Potentially, the following aspects of the voice interface could be tested:
1. Is it clear how to record and cancel voice recording? Is it evident that you can fast forward while listening?
We decided not to test these aspects because established patterns exist among competitors, and public messengers such as Telegram, WhatsApp, and Viber are used daily by our users. We didn't want to retrain people on how to use voice features.
2. Is it clear what actions are possible with voice messages?
There was no point in testing this aspect either because we had no intention of changing the pattern for invoking actions on messages to maintain a consistent experience with other message types. Therefore, we wouldn't have deviated from this decision anyway.
Results
First and foremost, the implementation of voice messaging increased revenue, as some clients immediately purchased the new feature, allowing users to record voices directly in Frisbee without switching to other platforms like Telegram or WhatsApp in some cases.
In addition, we have resolved the issue for potential customers who requested voice messaging during the product testing period.
A month later, we measured Adoption Rate and CSAT for voice messaging, which showed excellent results.
My Role
– Kick-off meeting with stakeholders and task understanding
– Discovery: interviews, desk research, and competitor analysis
– Formulation of hypotheses and Lo-Fi design
– Prioritization and scoping
– Presentation and defense of design to CPO and clients
– Hi-Fi design, requirement descriptions, and corner cases
– Creating animations and handing them over to developers using Lottie
– Design review before release, logging design bugs
– Monitoring metrics after release
– Product development after the launch of the first version