My goal was to understand why users were ignoring content and improve discovery without creating visual chaos.
We saw a high drop-off shortly after users entered the platform.
The problem was not that users could not access content. They could. But after logging in, they were immediately shown a large feed of videos with no guidance and no personalization.
From user interviews, I learned that people did not clearly understand where to start or which content was relevant to them. The platform had a lot to offer, but the first experience felt like a wall of content.
The onboarding became the first step in personalization. Instead of asking users to explore a large library on their own, the product started guiding them from the beginning.
The first screen should not show everything the product has. It should help the user understand what is useful for them first.
Users wanted to find specific videos, but the search was too easy to miss.
Instead of using search, they were manually scrolling through the platform and looking through many videos one by one.
This made discovery slower and increased the chance that users would leave before finding something relevant.
Search became a stronger part of the discovery experience. Users no longer had to rely only on manual browsing.
Popular categories also helped reduce the "empty search" problem by giving users ideas of what to explore.
Search only works if users can actually notice it. Sometimes the problem is not the feature — it is its visibility.
The homepage had a lot of videos, but most of them looked very similar.
I ran a 5-second test with the old design. After seeing the homepage for five seconds, users could not remember even one specific video.
This showed a banner blindness problem — there was a lot of content on screen, but nothing stood out enough to be remembered.
We didn't need more content. We needed to make the existing content easier to notice.
Instead of giving every video the same visual weight, I created more contrast between content blocks.
Consistency does not mean making everything look the same. A good system needs enough variation to guide attention.
The redesign moved the homepage away from one repeated visual pattern and towards a system where different types of content could stand out while still feeling consistent.
The wider product improvements contributed to a 17% increase in activation and improved D7 retention from 21% to 40%.
The new content system also gave the team a scalable way to work with thousands of videos instead of treating every thumbnail as a separate design task.
Consistency and repetition are not the same thing. A product can stay consistent without making every piece of content look identical.
The design system should create enough structure to keep the experience clear and enough variation to help users notice what matters.