DOU+ ONBOARDING FLOW OPTIMIZATION

At the time, the business faced a key challenge: encouraging more users to upgrade their accounts through DOU+. Significant operational and marketing resources had already been invested to drive this behavior, but the actual results fell short of expectations. As a result, the business began to shift its focus from purely operational tactics to improving product capabilities as a more sustainable growth lever.

Specifically, onboarding performance did not meet business targets. The expected trigger rate for onboarding tasks was over 50%, and the expected onboarding completion rate was also above 50%. However, actual performance consistently failed to reach these benchmarks, making it difficult to support ongoing business growth.

I started by analyzing existing data and found that the onboarding upgrade flow could be broken down into three steps: clicking the "Upgrade" entry point, submitting required information, and passing the review. The main drop-off occurred in the first two steps.

To identify the root causes, I reviewed all live onboarding flows using accounts with different statuses. Since user account conditions varied, there were four distinct upgrade paths in production. This review revealed several issues, including redundant steps, unnecessarily complex flows, and simplified paths that were not surfaced at the right moments. In addition, insights from customer support showed recurring user complaints: unclear upgrade value, confusion about account status, and a process that felt overly complex.

I started at the very top of the funnel. The entry page was trying to do too many things at once, which made the value of upgrading hard to grasp. I reworked the information hierarchy to clearly surface why upgrading matters, before asking users to take action. I also separated the paths for personal and business accounts to reduce early confusion and decision fatigue. To validate the changes, I explored and tested multiple content and layout variants, focusing on which version could most effectively encourage users to enter the upgrade flow.

For the Entry Page Optimization, the core issue was that the original page did not clearly communicate why onboarding mattered. The interaction and information structure made it difficult for users to understand the value of upgrading, which reduced their motivation to continue.

To address this, I designed two different content strategies. One used a high-information-density layout to clearly explain the value of onboarding, while the other compared user benefits before and after onboarding to make the upgrade more tangible. I tested these approaches across multiple experiment groups, updating the visual style from muted tones to brighter colors and introducing a black-and-gold theme to convey a more premium feel. Since user preference was uncertain, all variants were evaluated through A/B testing, with detailed results discussed in the next section.

To validate the final impact of the experiment, we decided to run online tests across multiple design variants. At the same time, we wanted to better understand how users respond to these different page designs and identify patterns that could inform future design decisions.

The experiment was structured as follows: