Increasing user retention by 7%
Talent seekers were taking too long to engage with their job openings after posting them, leading to a high dropout rate following their first publication. We identified the root cause as the complexity and unclear terminology of various platform features, which posed challenges for new users. To address this, we developed a self-guided tour that briefly explained the most important sections in the platform for the talent seekers. This initiative resulted in a 7% increase in user retention, a 14% improvement in positive user feedback, and a 37% completion rate for the tour.
Company:
Torre
Role:
Product Designer
Tools:
Figma, HotJar, Zoom, Notion, Metabase
Context
Torre is an AI-powered platform that connects job seekers with opportunities and helps talent seekers find ideal candidates by analyzing skills and preferences. For talent seekers, the following sections of the platform are particularly relevant:
Job post: View where talent seekers provide all the necessary details about the opportunity, including responsibilities, compensation, location, required skills, contract type, and more.
Pipeline of candidates: A board-style view of the recruitment process for a specific job opportunity. It allows talent seekers to review and organize candidate cards, from initial matches to potential hires.
Genome: A concise summary of candidate information, including skills, interests, professional experience, compensation preferences, reputation, and compatibility with the job opportunity.
Problem
We had two main problems related to talent seeker engagement:
The activation time (how much did the talent seeker took to interact with their job opening after posted) was too long.
The dropout percentage (talent seekers not returning after posting a first job) was too high.
We attributed these issues to some features or concepts that we had in Torre that weren't simple to understand for new users. Why?
Some concepts were different to the ones used in other job boards or complex to understand. For example, the genome, usually known as a profile, or the match and rank, an algorithmic tool that matches candidates with jobs using hundreds of factors.
In previous research we had received feedback that some of these concepts weren't simple to grasp at first and that it took some time to get used to them.
Solution
Step 1: Ideation
Together with the team, I brainstormed several solutions to address the issue, including:
A library of short educational videos.
Adding tooltips to clarify confusing sections.
A self-guided tour of the platform's main sections.
The only consensus at this stage was that the solution should be embedded in the platform to avoid relying on manual onboarding by the Torre team.
Step 2: Understanding the opportunity
I conducted research to better understand the challenges, which included:
Reviewing user session recordings to analyze how talent seekers interacted with the platform.
Conducting exploratory interviews with talent seekers who had dropped out after posting their first job.
Gathering feedback and reviewing previous reports about talent seekers' pain points.
This research highlighted a need for automation and in-product support to assist talent seekers in key workflows, such as posting jobs, managing pipelines, searching for candidates, and interpreting candidate profiles.
After aligning with the Product Manager and the CEO, we decided to move forward with the self-guided tour because:
Manual onboarding was not scalable.
It allowed us to track at which step talent seekers left the tour.
The tour's steps could be updated or adjusted as new insights emerged.
Step 3: Hypothesis
With the Product Manager, we hypothesized that the self-guided tour would:
Increase talent seeker retention.
Reduce activation time for talent seekers.
Improve positive feedback on the user experience.
To measure its impact, we planned to analyze retention, activation time, positive feedback, and the tour's usage and completion rates.
Step 4: Analyzing alternatives
We benchmarked onboarding flows and elements from platforms like Slack, Duolingo, Grammarly, Typeform, HubSpot, and Jellyfish to identify best practices.
We also evaluated whether to develop the tour in-house or use a third-party solution. After comparing providers (UserGuiding, Chameleon, UserPilot, and ProductFruits) on pricing, features, and integration capabilities, we opted for an in-house solution. This approach allowed for faster deployment and seamless integration with our design system.
Step 5: Design, specification, and deployment
I designed and developed a step-by-step self-guided tour to introduce users to key platform sections, including:
Job openings.
Candidate pipelines.
Genomes (candidate profiles).
To avoid overwhelming users, we included an option to snooze the tour, which would resume the self-guided tour at the same step when users revisited these sections.
Given that the tour relied on widely-used and proven UX patterns from top platforms, we deemed a usability test unnecessary. After development, we conducted a thorough QA process using feature flagging to ensure smooth transitions, accurate positioning of the suggestion boxes designed, and seamless functionality.
Key takeaways
A couple of months after deploying the self-guided tour, we analyzed the metrics defined in our hypothesis. The results were as follows:
Talent seeker retention increased by 7%, measured by whether users returned to the platform to post a second job.
The guided tour had a completion rate of 37%.
Positive experience feedback improved by 14%.
Activation time remained too variable month-over-month to attribute changes directly to this effort.
Based on these results, we decided to retain the self-guided tour as a core feature. Moving forward, our focus will be on researching and implementing strategies to increase the tour’s completion rate.