Our team working together in our research and iteration phases
Empowering amateur musicians, content creators, and businesses to make original music.
Amper is a cutting-edge, AI-powered music creation app designed specifically to streamline and expedite the process for content creators that have minimal experience creating music. Users only need to select a musical style, mood, and duration, and then Amper will create a bespoke piece of music in seconds.
As cool as this sounds, still retention was painfully low. Collaborating closely with cross-functional engineering and business teams, our overarching aim was to launch a new experience that would significantly enhance user retention by providing a user experience that surpassed industry standards. This was particularly essential as we were preparing for the imminent release of Amper's new subscription model. Our journey was a blend of data-driven insights, innovative design thinking, and extensive user testing.
Pen & Paper!
Role & Responsibilities
As the Product Designer and Product Lead on this project, my main responsibility was to ensure that our design decisions aligned with the user needs and business objectives. I was involved in every stage of the product development process - from initial research and ideation, through to design and testing. I took ownership of researching, planning, designing, and testing all interfaces for the Project Templates feature, and worked closely with the Engineering lead, and front-end, and back-end engineers.
Despite gaining considerable attention and press in major publications worldwide within just five months of its launch, Amper was grappling with a critical issue: our user retention rates were low, impeding our business growth and undermining our otherwise high ROI on customer acquisition.
Upon further investigation, we identified two main factors contributing to this issue:
The product experience was not intuitive for non-musical video creators. This presented a significant usability barrier and impeded smooth onboarding.
The product did not seamlessly integrate into the users' existing workflows. This friction meant that even if users were initially intrigued by our unique offering, they did not find sufficient value in incorporating it into their routine.
Although the Amper Music platform is built for all creatives who lack experience with music production, we investigated a particular interest among video content creators.
How might we enable emerging creators without music experience to efficiently find, evaluate, and edit music for their videos?
For our user research, we specifically targeted video creators without music expertise who had purchased stock music for a video project within the last three months. Our goal was to gather a deep understanding of their workflow, needs, and pain points. We collected extensive feedback through in-app feedback forms and video interviews, which were then synthesized into themes.
One of the main themes that emerged from our research was the mismatch between Amper's offering and the video creators' existing workflow. It became evident that to improve user experience and retention, we needed to adapt our product to better fit into the users' current workflow.
Amper had a limited offering of four genres, each with about ten music styles listed as text. This was restrictive and did not cater to our users' needs, particularly those who are visual and non-musical.
Our rendering process was time-consuming, requiring users to wait for up to a minute before they could listen to their first track. This impacted the user experience negatively.
While our AI-generated music delivered impressive results for some genres, it struggled with others. Genres like rock were challenging to render convincingly, leading to a varied quality of music output.
Though Amper offered unique music editing capabilities that stock competitors couldn't, the process was quite complex and required some level of musical expertise. Moreover, each edit necessitated a new render, adding more time to the process.
The ability to download individual tracks for each instrument was a strength, as it provided users with a high level of flexibility and control.
How did we stack up to the competition?
In order to fully understand the competitive landscape and identify opportunities for differentiation, we carried out a comprehensive competitive analysis. We synthesized our initial research findings and categorized features into core, preferred, and differentiating categories. This enabled us to create a competitive matrix that measured how Amper and our key competitors catered to these three types of user needs.
Our competitive matrix was divided into two sections: stock competitors on the left and algorithmic and AI competitors on the right. The Y-axis represented business success, providing a clear view of where Amper stood in relation to others in terms of business performance.
From this analysis, we learned that Amper outperformed its competitors in terms of differentiating features, generating considerable excitement about our product. However, we fell short when it came to meeting the core user needs, which are essential for ensuring a positive initial user experience and fostering long-term retention.
Based on these insights, we decided to prioritize improvements in two key areas: speed and accuracy. Let me explain what that meant:
In successful applications like Instagram, TikTok, or even Tinder, the speed of content delivery creates a "just one more" phenomenon. Users are enticed to keep engaging because the response is so rapid. This is the hook that kept users engaged with stock competitors. Conversely, slow response times lead to a negative user experience, which was a challenge with Amper.
High playback accuracy ensures that users get the right music options that truly meet their requirements. While accuracy doesn't need to be perfect, it needs to be good enough to keep users engaged and help them find what they're looking for.
Prototype & Testing
Our approach to improving the Amper app was grounded in an iterative process of prototype testing and refining, broken down into two main phases: the InVision Prototype Testing and the Closed Beta Launch.
InVision Prototype Testing
In this initial phase, we employed an InVision prototype to test the waters. We enlisted the help of ten video creators who were frequent purchasers of stock music but had never used Amper. We tasked them with specific actions and probed their expectations concerning the type of music they wanted to hear.
The results were largely positive: 9 out of 10 participants completed all the tasks, and their feedback provided invaluable insights. We learned about the features that resonated with our users, and the ones that fell flat. For instance, offering track suggestions for specific video use cases proved helpful to our users. However, we also discovered a significant hurdle we needed to overcome - the communication barrier between non-musical users and our inherently musical technology.
For the closed beta launch, we selected 100 Amper users who had rendered music at least three times within the past three months but had never exported their work. We chose this group specifically to test if the new features and improvements could convert these 'renderers' into 'exporters'.
The key performance indicators for this phase were:
Speed: We succeeded in decreasing the time to first preview from an average of 63 seconds to just 2 seconds. This dramatic reduction in load time brought us into a competitive speed range with stock music sources.
Accuracy: The weekly unique render to unique export ratio increased by 44%. This was a significant metric for us, indicating that more users were finding the music they needed quickly and were ready to export it for use in their projects.
Results & Outcomes
The iterative process of prototype testing, user feedback, and refinement led to substantial improvements in both user experience and business metrics. Here are some key performance indicators and business wins that resulted from this project:
Key Performance Indicators:
Monthly Rendering Retention: We saw a phenomenal 243% increase in monthly rendering retention. This surge suggested a significant improvement in user engagement and product satisfaction.
Weekly Unique Renders per Unique Export: This metric went up by 31%. The increase indicated that users were not only creating more music using Amper but were also finding music that was suitable for their needs, leading to a higher rate of exports.
Integration with Premiere Pro: This integration allowed video creators to easily access and use our AI-powered music directly within their preferred video editing software, thereby widening our user base and streamlining the user experience.
Integration with Reuters Web App: Our collaboration with Reuters further extended our reach, making our service available to Reuters' vast user base.
Subscription Model: After refining and enhancing the user experience, we successfully launched a subscription model for our paying customers. This marked a significant milestone in our monetization strategy.
Influenced Amper's Acquisition: Perhaps one of the most exciting outcomes of this project was that the positive outcomes of pivoting from start-from-scratch music creation to editable stock music influenced our acquisition by public stock asset company Shutterstock!
Overall, the results of this project marked significant milestones for Amper. The growth in user engagement and retention, combined with the successful subscription model launch, strategic integrations, and a successful exit points to a bright future for our AI-powered music creation app.