top of page

Boosting registrations for Stepstone’s AI Interviewer

Project scope

Team: Product manager, 2 product designers, 1 data analyst / prompt engineer, devs, 1 user researcher (me)
Timeline: 4 weeks 

Company: Stepstone

Methods: Usability testing, analytics, workshop

The challenge

Stepstone’s new GenAI-powered Interviewer had low conversion post-launch. The product aimed to drive registrations and recommendable profiles (i.e., profiles that can be recommended to recruiters), which it didn't deliver on.

​

The team wanted to understand where and why users were dropping off and what issues hindered them from converting.

The approach

I led a 4-week discovery and ideation phase that included:

  • Funnel analysis to identify drop-off points

  • 12 moderated usability testings with active jobseekers

  • A cross-functional workshop to translate findings into a prioritized roadmap

My impact

  • I identified key usability issues on mobile that were blocking users from converting.

  • Through a collaborative workshop, I helped shape the product roadmap and inform design changes that contributed to an 80% relative increase in conversion.

The process

1. Alignment and status quo

​​

Analytics revealed low funnel completion and notable drop-offs, especially on mobile compared to desktop. I partnered with the product manager and data analyst to find out where users were exiting and identify patterns in the funnel, which helped us form our first hypotheses.​
 

Additionally, most of the traffic came from mobile, but many steps in the flow weren’t mobile-optimized.

​

To better understand our users, I also requested additional context, such as common job titles being practiced, to ensure the research would reflect the actual audience landing on the page.

2. Usability testing​

 

I ran 12 moderated usability tests with users who were:

  • Actively job-seeking

  • Not yet registered with Stepstone

  • Struggling with interviews (our target use case)

  • Junior to mid-level job seekers, as these roles were most commonly practiced for.

​

Participants were recruited via an external platform and I targeted active jobseekers struggling with interviews, and not yet registered with Stepstone, ensuring alignment with the target audience.

We learned, that...

  1. The value prop wasn’t clear. Many expected a live coach or prep tips, not a mock AI session. 
     

  2. Friction on mobile was high. For example, uploading CVs was difficult, and a bug on Android blocked uploads entirely.
     

  3. Interaction mechanics were unclear. Participants didn’t always realize the mock AI interview session had started or that they were being recorded.
     

  4. The AI feedback was highly praised. Once users reached the actual session, they found the feedback helpful and actionable for improving their interviewing skills.

​

In short, the core product brought them a lot of value, but participants weren’t making it there.
3. Workshop

​​

My product manager had an off-site already planned to structure a potential roadmap, so we aligned in advance to carve out time for an ideation session during that.

​

Alongside my product manager, I co-facilitated an ideation session with designers, engineers, and the prompt engineer to turn these findings into action.​

 

Together we:

  1. Reviewed the relevant research

  2. Explored competitor flows

  3. Ran a brainwriting session to generate ideas for the mentioned friction points

  4. Mapped solutions by impact vs. effort to define an actionable roadmap

 

After the workshop, the designers took the lead on redesigning the experience, and the devs implemented several of the improvements, including a clearer landing page, simplified job input, and key bug fixes, among others.

The outcome

After the first set of changes, such as simplifying the job description input, we already saw a relative increase of 80% from the landing page to completing one mock interview session.​

​

Not all recommendations were implemented during my time at the company, so I can't speak to the final impact of it all.

Reflections and learnings

Pair data with user empathy

Funnel analytics showed where users dropped off, but usability testing helped us understand why. Aligning both gave us a much clearer picture.

​

Align with SEO/SEA teams early

The product was originally optimized for desktop, based on prior survey data, stating that the majority of job seekers prepared for interviews on desktop. But SEO and SEA efforts brought in mostly mobile traffic, which created a mismatch we could have anticipated with closer cross-team alignment early on.
 

Involve developers in the process

Bringing devs into ideation sessions surfaced creative ideas we might’ve missed otherwise. Their solution-oriented mindset helped bridge the gap between insight and implementation.

bottom of page