I was in charge of rethinking the most relevant area of our web product ⏤ the Hiring Funnel, from the ground up. Once I understood the current problem, and analyzed our users’ behaviors, I ran the initial explorative research with our users through 1:1 interviews and creative brainstorms. With the learnings I started designing, and proposed an ideal version (to serve as the vision) together with an MVP version that we could easily test against the current situation at the time.
The company had been trying to improve the Hiring Funnel since before I started and they were starting to see diminishing returns from every test they were running.
We needed to re-think the Hiring Funnel from the ground up to enable growth.
I analyzed our user’s behaviors through Hotjar ⏤a tool that allows you to record your user’s actions on your website⏤ and learned they tend to follow a non linear path. They move back and forth in the funnel, without converting.
I wanted to learn what was happening behind the scenes, why were they moving back and forth? What were they really looking for that we were providing on their first try?
So I set up a few 1:1 interviews to learn more.
First I tried to understand more about who they were and what were their motivations and assumptions. Then I asked them to go through a card sorting exercise where they had to rank all the elements that make up their decision of either booking a task or not.
Once I had enough information, I gathered all the stakeholders and ran a brainstorm around all the possible crazy ideas for this new, improved Hiring Funnel.
I started wireframing a new flow, starting from the homepage.
I went for a solution that puts front and center the information that ⏤from what I learned during research⏤ was most important for our clients when booking a task.
The biggest change that I introduced, had been entirely removing one step from the flow. I wanted users to go straight from homepage to the list of Taskers, as we’ve learned “choosing the right Tasker” was the most relevant decision they had to make in order to book.
Once the wireframes were approved, I moved to visual design.
I proposed various options and ran few usability tests to understand which one would have been more attractive to users.
Thanks to a series of design reviews that followed, I was able to propose an ideal version of what this new Funnel could have looked like, which you can find below.
• Search parameters at the top so the user always knows what list they’re looking at
• Show all available Taskers and allow the user to filter down based on their preferences
• Add more filtering options on the left side, most notably, a price graph
• Allow users to favorite Taskers so that if the one they choose end up not being able to do the task, they have already selected a few that they liked
• Give users ability to choose a time-window (i.e. “Next Three Days”)
• Reduce size of Tasker cards so we can show more than one above the fold
While I was creating variations to this design, the engineering team just completed re-writing our website to React.js which now enabled us to move faster and create re-usable components.
So I decided to lay the ground for our future Design System. Regardless of the result of this launch, creating these components would have been an investment for the future. In fact, we still use them today.
Minimum Viable Product
After agreeing on the vision we decided to reduce the design to a version that would allow us to validate the assumption that showing the list of Taskers with filter is a successful idea.
I created a prototype to make it easy for stakeholders to understand what we wanted to test.
Once the MVP was ready we launched the test to half of our users and started tracking key business metrics, with a focus on conversion.
To everyone’s surprise conversion was slowly decreasing.
After implementing small improvements, the team decided to pull the test a declare it failed.
This was my first project at TaskRabbit and it allowed me to learn immensely. Here are just some of the lessons I learned:
1. Have a phased approach when testing.
We tested launching new features, and optimizing of existing features together. This resulted test pollution and didn’t allow us to clearly see what was affecting the drop in metrics.
2. There’s good friction and bad friction.
I removed a step in the funnel were users had to describe their task before seeing Taskers. That friction increased user engagement, resulting in users that are more invested and more willing to move forward.
3. Be mindful about what you remove, no matter how much you add.
Regardless of how many features we added, I removed a key element. I removed the review from the Tasker card in order to make it shorter.