LANDR · 2021 · Senior Product Designer · ~3 weeks · Shipped · Measured
Pros churned because LANDR's mastering gave them almost no control. After seven interviews with churned Pros, I designed a separate revision tool that asks plain questions, instead of complicating Mastering with technical jargon. One failed test and one extra week later: 80% task success and $147K in incremental revenue.
For producers who know what they want, LANDR's one-click mastering felt too basic. After a few disappointing masters, many considered other tools. Churn and weak conversion to paid plans followed.
A churn survey, then seven interviews with Pros who had churned. All seven wanted a full plugin's power with much easier usability. Five had loudness problems, five had clipping, four wanted finer EQ.
Stakeholders wanted an extended version of Mastering. The PM and I said no: new users and Aspiring Musicians loved its simplicity. Revisions became a separate Pro feature, linked to the original master.
Technical copy, confusing concepts, no audio while revising. Leadership wanted to launch anyway. I asked for one more week. Version two asked plain questions with opinionated answers and let people listen while revising: 80% task success on the same panel.
The subscription message leaned toward a dark pattern. I moved the upgrade prompt to after "Revise master", once people had invested, and got an A/B test planned before I left.
On the Mastering Pro journey, the lowest point came after the master: a Pro, consistently disappointed by the result, wanting more control. The original product felt too basic for people who know what they want. After repeated disappointments, many considered switching to other mastering tools.
Pros generated millions a year, and this pain point drove churn worth hundreds of thousands in lost revenue. With thousands of potential Pros converting to paid plans each month, weak conversion missed hundreds of thousands more. The goals: more plan upgrades, less Pro churn, better engagement and satisfaction.
A churn survey to Pro subscribers confirmed most of them were Pros, and that they wanted more from a mastering tool. Then I interviewed seven who had churned, combining the Five Whys for root causes, motivational interviewing for goals, and Jobs-to-be-Done for the outcomes they needed.
Granular control over things like EQ and sound types, in an interface that stays intuitive, fast, accessible and easy to navigate. Prototype and test until success passes 75%. My hypothesis: Pros know mastering well enough to adjust a track that's already mastered.
Key stakeholders suggested an extended version of the Mastering product. The PM and I rejected it. The original was so successful it had made LANDR a flagship, and we didn't want to disrupt it for new users or Aspiring Musicians, who valued its simplicity. We built a separate Pro feature, still linked to the original.
A professional mastering tool is technically complex. Luckily, the PO was a seasoned mastering engineer. Through several ideation and wireframing sessions, we balanced technical needs with usability, and built the first prototype.
Five tests on the same panel of users, to avoid delays. Task success: 20%. People were confused by some concepts and struggled to choose settings, which went against my hypothesis. The copy was too vague. And they wanted real-time audio while revising, which the servers couldn't do.
Leadership wanted to ship after the first round, despite the results. I advised against it: frustration for Pros, damage to LANDR's reputation, and little chance of meeting our goals. Improvements were cheap. I convinced them to extend testing by a week, and they agreed.
An introduction to reduce confusion. The ability to listen to the master while revising, so people judge by ear instead of by memory. And questions with clear answers: "How is the loudness? Too quiet / Good / Too loud." "The lows are: Good / Boomy / Thin." Familiar language, plus tooltips for the theory Pros sometimes lacked. People chose with more confidence and less hesitation. 80% task success.
Non-subscribers saw a message saying a plan was required. I asked the design team for a less intrusive pattern, since it leaned toward forcing people to subscribe before trying the feature; the team decided it wasn't a priority. So I worked with product to show the upgrade prompt only after people click "Revise master", once they've invested. Before I left, it was on the roadmap, with an A/B test planned against a hidden paywall. Close to finishing, people want to finish.
80% task success, $147K in incremental revenue, 9% more plan upgrades, 5% less churn, and 11% of Pro mastered tracks revised at least once. Feedback after launch confirmed Pros were eager for a feature like this, and that it played a key role in their decision to stay with LANDR or subscribe.
Interviews told me Pros knew mastering. The test showed what they actually did. Observation after interviews is how you validate an insight: what people say isn't always how they behave. One note: the live panel was changed after I left, without my knowledge. What's shown here is my version.
Research, interaction, UI, copy. With the PM, a PO who was a mastering engineer, engineering and the design team.
Real-time revision audio was impossible on the servers. The live UI was changed after I left; the v2 shown here is my design.
Observation beat interviews: what people said wasn't always what they did.
Next case: The wait became the product