The onboarding was not too long. It showed value too late.

People kept taking screenshots of one part of the product: the style analysis. They shared the result because it gave them a specific description of how they dressed and something recognisable to react to.

The onboarding did not lead with that. It asked for a long sequence of setup information, introduced several features and ended on an empty Saves tab. The product had something people valued, but it arrived after most of the asking was done.

The old flow asked before it gave

The previous sequence stretched across 13 screens. Name, handle, shopping preferences and feature explanations all appeared before the person had seen a result made from their input.

A selfie also appeared without enough context. It looked like another requirement in a setup process rather than the input to an immediate, personal result. Even someone willing to complete the flow had to trust that the useful part would come later.

Move the proof to the front

The new sequence began with one lightweight preference and moved quickly to the photo. The interface explained what the image would be used for, then showed the analysis before asking for the remaining profile details.

That changed the meaning of the later questions. A name and handle were no longer payment demanded before entry. They completed a profile whose value the person had already seen.

The wait needed to remain part of the story

The analysis took roughly 15 seconds. Hiding that behind a generic spinner would have broken the connection between the photo and the result, so the processing screen kept the image visible and described the work in progress.

The goal was not to make the wait theatrical. It was to reassure the person that the system still had their input and that the next screen would answer the question they had just been asked to invest in.

The reveal needed pacing

The result was split into a short sequence rather than one dense report. It moved from broad observation to specific traits and finished with a compact style label that could be saved or shared. Each screen had one job and enough time to be understood before the next appeared.

The rest of the product had to continue the same promise

After onboarding, the first tab changed from an empty collection to a populated discovery view. Recommendations followed from the preferences and analysis the person had just seen. Saves remained important, but they no longer had to explain the value of an empty product on their own.

Most of the underlying capabilities already existed. The redesign changed their order and the responsibility of each screen. The product asked for less trust at the beginning, showed a real result sooner and made the remaining setup feel connected to something the person had chosen to keep.