ToptalTalent onboarding

Rethinking a 3.5-month onboarding flow

Why so many applicants never finished

Toptal’s talent onboarding took applicants anywhere from three weeks to five months, averaging nearly three and a half. Many never finished.

Soon after joining the team, I worked with a small group to find out why, then redesigned the flow around what applicants actually needed.

Role
Senior UX designer
Team
Shared researcher, engineering, and operations partners
Timeline
2018
Status
Shipped
First step of the redesigned flow: a four-step progress bar, a welcome message, a green Schedule Call button and a tips card.
A screen from the redesigned flow: the English call step, with all four stages visible up front.

Why so many applicants never finished

Toptal is a freelance network that connects engineers and designers with companies. Before talent can take on client work, they go through onboarding: screening calls, forms, a portfolio review, technical interviews, and a test project. When I looked, that took anywhere from three weeks to five months, averaging nearly three and a half.

So many applicants didn’t finish. I joined as the team’s second UX designer and picked this up as one of my first projects, working with a small team to find out why and to design something better.

I did the UX and UI design for the new flow, and worked with a researcher who was shared across the whole design team.

Going through the process myself

I didn’t have the product background others on the team had, so I started by learning the process firsthand. I asked to sit in on every onboarding step that involved Toptal staff: the English call, both technical interviews, the portfolio review, and the development test. I also went through onboarding myself, repeatedly, to experience each situation applicants ran into.

Then I signed up with the competitors applicants had told us they turned to instead of Toptal, across several talent groups. I compared how easy each was, what it required, how much it made me trust the company, how long it took, and how many steps it involved.

The pattern was clear. The flow had been built for the ease of back-end development, at the expense of the applicant. Applicants answered dozens of questions, some of them more than once, and completed technical interviews and a test project.

The tradeoff was patience. I didn’t draw a screen until I understood why the flow looked the way it did.

The flow made things easy for the back end, at the applicant’s expense.
Eight competitors' sign-up pages side by side, each labeled with the company's name.
Sign-up pages from the eight competitors I compared Toptal against.
A Sketch file open to its Competition page, with rows of small competitor screens in clusters on a dark canvas.
The competitive analysis in Sketch, with screens captured from each competitor’s flow.

Counting the friction, and naming its cause

To make the problem visible, I mapped the two main paths applicants took. Toptal’s system sorted them into high, normal, and low priority flows with a 16-question classifier. Even the best case, the low-friction flow, took 8 screens, 7 emails, 27 form fields, and 4 phone calls. The high-friction flow, for “normal” priority applicants, took 9 screens, 7 emails, 38 form fields, and the same 4 calls. Applicants on that path were far more likely to give up, usually around seven weeks after applying.

The biggest pain points were the classifier, the English screener (a required call whose main purpose was never explained to applicants), portfolio voting, and the technical interviews.

The friction wasn’t purposeful. Toptal had grown quickly without enough UX involvement. So before reworking the whole journey, we encouraged engineering to take some immediate steps we believed would produce easy wins. The tradeoff was patience again: easy wins first, the full redesign second.

The friction wasn’t purposeful. Toptal had simply grown too quickly without enough UX involvement.

A wide flow diagram titled Low-friction flow: one left-to-right row of steps, from Home and Sign-up to Dashboard.
The low-friction flow mapped step by step, from sign-up to the applicant’s dashboard.
A wide flow map on a white canvas, from Talent on the left to Value on the right, branching into three horizontal paths.
The full flow map, from Talent on the left to Value on the right, with paths that branch by talent type and priority.

Letting applicants finish in any order

With the information we actually needed defined, I explored different ways to group the questions and tested dozens of layouts. This overlapped with BASE, so I could swap components and patterns quickly. I took the directions back to internal operations staff and recently hired talent for feedback.

The final design made it clear what was expected at every point. Applicants could fill out forms in any order and come back later. Before, it wasn’t obvious whether their work had been saved. Instead of involving a screener at four different points, I proposed combining the calls into a single step, and the designs dropped optional fields and unnecessary emails. There were also small moments to learn about Toptal along the way, without being heavy-handed.

The old flow put back-end convenience first. The redesign put the applicant first, and that meant asking engineering to accept some change.

Applicants knew exactly what was expected of them at all times.
A large canvas of about seventy small Toptal screens in columns, mostly white or blue pages with forms and cards.
A sample of the layouts and question groupings explored along the way.
The Portfolio Review step: a progress bar, upload instructions, a green Upload CV and Portfolio button, and a tips card.
The Portfolio Review step of the final flow, with the finished English call checked off.

A simpler flow that shipped in 2018

Advocating for applicants at a pace engineering could accept

This project taught me to be an advocate for the end user while staying aware of the team that has to build the change. The engineers I worked with were excited to make improvements, but they were wary of change, so I paced my suggestions and later designs to match.