After every session
A written debrief, the moment you finish
Three parts, every time. What worked, what to sharpen, and one thing to try next session. Each one points at something you actually said. No score, no grade, no ranking against anybody else.
Every debrief draws on the same coaching themes, shaped by real interviews, and on nothing but what you said in your session. See how your coaching works →
DashboardProgressResourcesSettings
New SessionWhat worked
Structure & Communication
You said up front that you'd cover the listener you were designing for, where discovery breaks down for them, what you'd build, and how you'd know it worked. Then you signaled each transition, so the reasoning was easy to follow even when you were thinking out loud.
Assumption Clarity
When you weren't sure how much listening history Spotify surfaces to the recommender, you named it as an assumption and kept going instead of stalling on it. You also flagged which of your assumptions mattered most to the answer.
What to sharpen
Trade-off Awareness
Your recommendation leaned on inferred intent, and you described what happens when the inference is right. Next time, say what happens when it's wrong: what the listener sees, and how confident the system needs to be before it acts.
Measurement Thinking
You named saves and completion rate at the end. Next time, bring measurement into your framing, and pair each metric with the thing you'd watch to make sure it isn't being gamed.
One thing to try next session
You raised measurement right at the end, after the solution was already settled. Next session, name the one number that would tell you the problem is solved before you design anything, then check each idea against it as you go.
Anything you'd add to this session? (optional)
How did your overall approach feel? What would you do differently?
What worked
Structure & Communication
You said up front that you'd cover the listener you were designing for, where discovery breaks down for them, what you'd build, and how you'd know it worked. Then you signaled each transition, so the reasoning was easy to follow even when you were thinking out loud.
Assumption Clarity
When you weren't sure how much listening history Spotify surfaces to the recommender, you named it as an assumption and kept going instead of stalling on it. You also flagged which of your assumptions mattered most to the answer.
What to sharpen
Trade-off Awareness
Your recommendation leaned on inferred intent, and you described what happens when the inference is right. Next time, say what happens when it's wrong: what the listener sees, and how confident the system needs to be before it acts.
Measurement Thinking
You named saves and completion rate at the end. Next time, bring measurement into your framing, and pair each metric with the thing you'd watch to make sure it isn't being gamed.
One thing to try next session
You raised measurement right at the end, after the solution was already settled. Next session, name the one number that would tell you the problem is solved before you design anything, then check each idea against it as you go.
Anything you'd add to this session? (optional)
How did your overall approach feel? What would you do differently?