I spearheaded the biggest redesign the company had ever gone through in order to solve a critical business problem: more than 75% of our users were dropping off.
Overpass provides a platform for freelance call centre contractors to create profiles in order to be hired by small businesses. Once a contractor creates their profile, they are required to submit the profile for review by our Talent Advocates. In January 2019, one of the biggest challenges the company was facing was that the majority of contractors weren’t submitting their profiles for review, leading to an unsustainably high drop off rate. This meant that even though many qualified contractors were signing up with us, they weren’t seeing the value we provided before they dropped off. It also meant that we were losing out on quality talent joining our marketplace.
We suspected that the cumbersome process of signing up for our marketplace was causing users to drop off before they submitted their profiles. In order to remedy this, we embarked on an initiative to completely redesign the contractor onboarding experience.
Contractors were signing up and dropping off at a very high rate. We decided to overhaul the registration and profile creation process to reduce the drop off.
I started out by gathering analytics data about our current contractors’ sign up behaviour. We learned that of the contractors that sign up, less than a quarter of them submit their profile. Of the contractors that submit their profiles, less than half are approved by the Talent Advocates. This confirmed our suspicions that users were dropping off before they complete their profiles.
We learned from our clients that the most important thing they look for on profiles is contractors’ previous work and industry experience. Through our research, we discovered that only 40% of contractor add previous work experience to their profile, and that the majority only add one past position. Our goal was to revamp the work experience interaction to make it easier for contractors to add their past experience, and make the work experience section more visually prominent in the profile so that clients could view that section more easily.
From this research, we confirmed our hypothesis that we needed to improve the sign up flow for contractors, improve the interactions for filling out various parts of the profile, and update the UI of the profile to match our new rebrand.
I started out by taking a critical look at how I had designed the original profile to determine which sections needed the most improvement.
In order to gain a deeper understanding how our contractors interact with the profile, I organized trips to our satellite office to interview the Talent Advocates, Sales and Customer Support teams. Through our discussions, I learned that when users were asked to enter free-form text, either in the bio section or the job description section, they would often leave it blank or write things that not only were irrelevant, but were actively hindering their chances of being hired. Even when the Talent Advocates explained the importance of the bio and what they needed to write, few contractors actually made changes or improved what they originally wrote. As a result, we decided to completely get rid of the free-form bio section, and significantly reduce the importance of the job description section.
I also learned from my interviews that numerous contractors would select many skills and industries when they had no experience in any of them. We decided to make sweeping changes to the way skills and industries are added to the profile.
In order to ensure that contractors actually had the skills they claimed and had worked in the industries they selected, we moved these sections to the work history and associated them with individual past experiences.
We also decided to require users to select a “Job Position” from a list. This enabled us to use this data in our matching algorithm, used to show clients relevant contractors. It also had the benefit of subtly reinforcing to the contractors what kind of experience is required to find a job on our platform.
We ask contractors what jobs they are interested in so that we can match them with clients. Similar to skills and industries, we found that contractors were selecting many jobs when they didn’t have any prior experience in those jobs.
In the new profile, we decided that this was important enough to warrant its own section. I took advantage of the increased space to describe to the user how the section works. We knew that contractors who already had experience in a certain job would perform better; at the same time, we didn’t want to discriminate against contractors who were just starting out and didn’t have enough experience. Instead of only letting users select jobs they have added in their Experience section, the compromise was to show a biased dropdown menu, with the jobs that they have added in their Experience section at the top, and the rest of the jobs they don’t have experience in at the bottom.
We also wanted the contractors to select one industry that they want to “specialize” in, which meant an industry that they already have added an experience entry for, and now want to find a job in. We restricted this selection to only the industries they have entered in their Experience section. This ensures that the contractors already have experience in that industry, which is one of the most important things a client needs.
When a desired job position and industry specialization is selected, the system automatically calculates how many years experience the contractor has in them, based on the Experience entries. We display the years on the client view of the profile.
In the original design, I wanted the users to review their profile before submitting it for review. My assumption was that users would want to check their work and make sure their profile was the best it could be before having it reviewed for approval. This assumption turned out to be wrong.
In session replays, I observed users skipping steps and putting in the minimum required amount of information before they submitted their profile. I believe this was because we asked for a lot of information that was tedious to enter, and we did not do a good job explaining to the user why the profile information was important. Therefore, users did not know what they needed to have a good profile, so they took the path of least effort.
To solve this problem, we decided to remove the manual submission requirement, and instead to automatically submit the profile when it has enough required information filled out. We would make it obvious to the user what was required and whether they had completed it via a dynamic checklist. Once the checklist is complete, the user is not required to take an extra, arbitrary step to submit their profile. We hypothesized that this would be one of the most significant factors in improving profile submission rates.
Our UI Designer created profile designs based on my wireframes and the style guide we were jointly creating.
The new profile interactions and design should improve the ease of filling out the profile, resulting in a better experience. The updated UI should give a sense of cohesion and continuity across the system. If the profile is easier to fill out and a more pleasing experience to interact with, the drop off rate of contractors should decrease.
These results should be verified by usability testing, which I outline in Proposed Usability Research.
In the original sign up flow, after registering we immediately take contractors through a walkthrough to help them fill out their profile. I originally designed it this way because we did not have anything else for the contractors to do on the platform, except to create their profile.
My assumption was that users would do anything to get a job, including filling out tedious forms with mass amounts of information. Through my research, I discovered this was not the case. I watched many session replays of contractors going through the sign up flow, and I realized that we were asking for too much input up-front from the user without showing them the value.
For the redesign, I decided to scrap the walkthrough completely. After registering, the contractor is taken to a new dashboard where they are able to see jobs they could apply to, once their profile is approved.
I designed the dashboard to show the contractors the most important things they need to know. In addition to the available jobs, they will see a checklist to tell them what to do next. The main call-to-action is to fill out their profile.
The dashboard would dynamically change based on where the users are in the process. Once their profile is approved, the checklist disappears and the “Apply” button on jobs will be enabled.
When a contractor is hired, their dashboard will display stats and information about their current job.
When our UI Designer started applying the UI to my wireframes, I realized that we were displaying too much information on the dashboard, with the potential to overwhelm the users. I worked with him to reorganize the dashboard to become more streamlined.
Showing users what they can do on the platform immediately after signing up should increase interest in the platform and reduce the drop off rate. Instead of taking them through a walkthrough, we allow users to fill in any part of their profile in any order, which should make them feel more in control.
These results should be verified by usability testing, which I outline in Proposed Usability Research.
In order to verify our hypotheses before spending development time building our new designs, I created a research plan outlining everything we want to learn and how to test it. I decided the best approach would be to conduct usability tests on the new sign in flow and profile interactions.
From these usability tests, I wanted to learn how users navigate our new designs, which parts of the UI they interact with, and which paths they take to accomplish their goals. In particular, I wanted to see whether it was clear to the user what they had to do at each step, whether they understood the process of creating a profile and having it reviewed, and whether it was clear to them when their profile had been automatically submitted.
The project is currently ongoing and the usability research is being conducted. Pending the results of the usability research, some UI or UX might need to be updated to ensure we are achieving our business goals and creating an enjoyable experience for our users.
This endeavour represents the largest redesign the company has seen since launch. Having designed the original system, it was exciting to see how my designs initially performed, then identify areas for improvement, and now back up my new design decisions with data. This approach represented a shift in the way we had previously designed at Overpass, which was to move fast and release often. I am proud to have led the charge to a new way of approaching design problems, by taking the time to do in-depth research on how users use the system and to hear from all the involved stakeholders.