Spotify had a problem.
They built a new feature powered by machine learning. It would show users their current favorite songs and playlists on the Home tab.
Simple concept. But nobody could agree on what to NAME it.
So they did what every tech company does: A/B testing.
Test 1: Failed. Inconclusive.
Test 2: Failed. Inconclusive.
Test 3: Still nothing.
The data couldn't tell them which name users preferred.
Here's where most teams give up or just pick randomly.
Spotify's UX research team tried something different:
→ Stopped looking at numbers
→ Started talking to actual humans
→ Asked people how they FELT about different options
→ Noticed something: users lit up when they saw personalized greetings like "Good Morning!"
They weren't responding to the NAME of the feature.
They were responding to feeling SEEN by the app.
Final decision? They added personalized time-based greetings throughout the interface.
The lesson here is simple but brutal:
Data shows you WHAT is happening.
Conversations show you WHY it matters.
Most designers worship A/B tests like gospel. But numbers can't capture emotion, context, or the "why" behind user behavior.
You need both:
• Quantitative data for patterns
• Qualitative research for meaning
Stop treating user research like a checkbox. Start treating it like a conversation.
Because sometimes the answer isn't in your analytics dashboard.
It's in the words your users choose when they talk about your product.
#UXDesign #UserResearch #ProductDesign #Spotify #DataVsIntuition