A Tale of Two Users.
We built a swipe to work app in Miami. I worked on the brand, the app… and everything in between. It's the project I think about first when I talk about my work.
The job search is clunky
I personally remember filling out an application for a job lifting boxes at the back of a warehouse when I was in high school only to get a response halfway through college (it was a no).
With , I was part of an ambitious project to redesign and modernize the way we find work.
Many pots on the stove
I came on as a product designer, but knew from the interviews that I'd be wearing a handful of hats.
I worked on the design of Fortuna from early 2022 to fall 2022. I initially collaborated with another designer before taking ownership of product design across the job seeker and employer experience.
There were just a few of us. I worked alongside that other designer, a couple software engineers, and a product manager who doubled as our co-CEO. Fortuna was an early stage startup, so roles bent and blurred. We all stepped outside our job description to keep the product moving.
launched in Miami in summer 2022.
Our North Stars
Our high level goals for the app were to:
- Drive the cost of every interaction towards zero.
- Automate the work so every tap is reserved for a decision.
- Keep the experience on one distraction-free path.
- Give users speed as a product; hired in hours, not weeks.
Field interviews
I partnered with our other product designer Yvonne, to explore how job seekers go about applying for work. We spoke to seven job seekers based in Miami and drew up Sarah, a persona based on our collection of insights.
Name: Sarah Anderson
Age: 31
Job: Unemployed
Located: Chicago, IL
User-persona situation:
Sarah is a 31-year-old administrative assistant. She worked as a receptionist for the last 4 years but recently lost her job. Sarah is a fast learner person who can adapt to different work enviroments. As a people person, she has excellent communication skills, and is always ready to face the day with a positive attitude. In fact, in her last job, she was recognized as the smiliest worker.
Sarah is a single mother of a 2-year-old girl and lack of time is a daily worry as she is currently paying for a babysitter. Finding a job within the neighbourhood would mean spending more time with her daughter. She wants to get a gig as fast as possible while looking for another job.
Situation related needs:
Sarah needs to find a job as fast as possible given the fact she is a single mother. In the meantime, getting a gig seems a really good idea.
As she wants to spend more time with her daughter, Sarah needs to find a job that’s not far from the neighbourhood.
Goals
Being a present mother for her daughter
Economic stability
Needs
Find a job near her home as she needs to be available in case of an emergency
Have an assured income while looking for a long term job
Find a short term job fast
Having free time to spend with her daughter
Frustrations
Don’t know where to start searching and feel overwhelmed with all the different job searching apps.
Searching in conventional recruiting apps takes her too much time as she needs to think and search for the specific type of job that she needs
Apply for each job and waiting to hear back from them makes her anxious as she needs an income as soon as possible
Attending to an inverview means she needs to pay for a baby sitter while she is gone, and she doesn’t have the income to cover that.
Sarah didn't think the job applications were horrid. They just assumed she had time to search when she didn't.
Seeing Sarah
If you really get into the nitty gritty of what Sarah needs, it isn't really a job.
The jobs? They're out there. But sitting between Sarah and a job are lengthy commutes, childcare stuff, a carousel of applications, a waiting game (that may all lead to nothing). It all really just boils down to friction.
Sarah just needs to reduce the friction between her current reality and the life she wants to build.
I feel like I was finally seeing Sarah's situation, she changed the way I wanted the product to work. Fundamentally, Fortuna isn't just job matching. It's more of a means to reduce the friction between someone's current reality and the life they want to build.
So the challenge is not to just “help Sarah apply to jobs”, but rather “how do we reduce the cognitive, emotional, and logistical costs of getting back to work?”
How do we reduce the cost of getting back to work?
Running on a hamster wheel in a rat race
We expanded Sarah's persona to get a clearer picture of Job Seekers in Miami. Job Seekers were vexed by the apply ➡️ wait ➡️ hear nothing ➡️ apply again hamster wheel.
Searching for work had become work itself. Every step demanded something from the Job Seeker: attention to search, judgement to filter, patience to wait. So basically by the time someone finally reached an application, they'd already spent hours working just to get there.
We didn't just want to make applying faster. We wanted Fortuna to make the work of looking for work feel lighter.
How might we make job seeking feel less like work?
The answer started with making job applications cost next to nothing.
Two sides, one app
Fortuna was a marketplace, so it had two sides: job seekers looking for work and employers looking for help. They met in the same matches, but almost everything else was different: the goals, the screens, and the story behind them.
We'll start with the job seeker. That's the side with the swipe, the onboarding, the bio and the video resume, and it's where most of Sarah's friction lived.
Tinder's spine
Job applications are full of small costs: read a job post, hit apply, reroute to another site, fill out another form. Individually these costs are trivial, but together it's enough to spur a job seeker from submitting an application.
Fortuna had to cost one gesture instead of several. Swiping already had a familiar mental model, so we hinged the user flow on it.
Swipe right to apply, and swipe left to dismiss.
Swiping error
The swipe interaction was familiar, but its outcome wasn't. When a job card disappeared, there was no immediate feedback to indicate whether the action had been an application or a dismissal. The interaction looked and felt like something users already understood, but without a perceivable confirmation, it broke the mental model they brought to it.
No interaction confirmation = confused users
Wait, come back
Testers were swiping through listings so fast that they wouldn't realize they skimmed past a potentially interesting job until it was too late.
Users were swiping wayyy too fast
Fortuna was a young pup back then, and every listing came from us knocking on doors around Miami (and it probably would be that way until the app gained some traction). So when users flew through the feed, they could run out of jobs in a single sitting before locking in and looking with more intention.
We'd spent weeks stripping friction out of Fortuna, but that made it easy to get lost in the sauce of the swipe, so to speak. Driving the cost towards zero had cheapened the interaction. The psychological delta between a thirty minute application process and a one second swipe was large enough to stop users from caring.
The psychological delta between a 30-minute application process and a one-second swipe was large enough to stop users from caring.
My friend Athena brought up Chesterton's Fence, a concept that warns against eliminating a rule or tradition before understanding why it was established in the first place.
Was there a reason the fence was up? We removed the friction fence because it drove the interaction cost up. But some of that friction had been quietly doing a job. It pushed people to pause long enough to make deliberate decisions.
So we'd removed the effort, but we'd also removed some of the intention.
What we shipped
Testing showed that reducing the number of steps did exactly what we wanted: users could move through opportunities quickly, without the usual application fatigue. But that efficiency came with a tradeoff. When the cost of a decision approached zero, some decisions started to feel disposable. Users would swipe quickly, sometimes realizing only afterward that they had passed over a job they were interested in.
The undo interaction became a small counterweight to that speed. It gave users a way to recover from a mistake, but limiting the number of undos available each day meant that reversing a decision still carried some weight. It was a small interaction, but it helped restore something we had accidentally designed away: intentionality.
The job seeker flow
fillerWith the swipe settled, here's the whole job seeker side, from opening the app to landing an interview. Most of it was built to get out of the way.
Step 1
Getting in
Splash → job seeker or employer → sign up → phone verification → location
Sign-up asked for only what we needed: a name, a birthday, an email and a password, then a phone number and a quick verification code.
Location did the most work. It decided which jobs you'd see — and, outside Miami, whether you could use Fortuna at all. Being a young pup meant one city, so anyone further afield landed on a waitlist instead of an empty feed.
Step 2
Three screens of setting expectations
Complete your profile → record your video resume → swipe to apply → I'm ready
Before the first card, three quick screens told job seekers what to expect: fill out a profile, record a short video, then swipe. No surprises later about why we were asking for a video.
Step 3
The bio
Name, location, max distance → industry, availability, wage → education, about me, attributes
The bio was the job seeker's side of the match. Distance, industry, full or part time and a minimum wage on a sliding scale decided what showed up in the feed; education, an about me and a handful of attributes — dedicated, loyal, punctual — were what employers saw.
Step 4
The video resume
Tips → camera and mic → record → play back → confirm and upload → uploaded
This was the heart of the job seeker side. Instead of a cover letter, a 90-second video introduction: a chance to stand out for people whose best qualities don't fit on a resume.
Most of the design went into making it less intimidating — tips up front (somewhere quiet and well lit, a plain background, no sunglasses or hats), one big record button, a playback before anything is sent, and a clear message while it uploads so no one closes the app halfway through.
Step 5
The feed, and hearing back
Home → swipe → my applications → employer interest
Then the feed — the swipe chapter above, in a sentence — and the part after it: keeping track of applications and hearing back from employers. To fill in from the next part of the flow.
From the job seeker to the employer
Fortuna had two users, but they were trapped in the same system. Job seekers were spending hours searching and applying with little feedback. Employers were spending hours sorting, interviewing, and coordinating with candidates who might never be a fit.
If Fortuna was going to make hiring faster, we couldn't optimize only one side. We had to remove work from both.
Meet Jonathanneeds a title
On the other side of the marketplace was Jonathan, a 38-year-old grocery store owner in San Francisco. Jonathan didn't need to hire constantly. He needed someone dependable who could step in when the store got busy: running deliveries, picking up supplies, cleaning, or covering whatever needed doing that day. His problem wasn't a lack of applicants. It was that finding the right person required more time than he had. Sorting through candidates, figuring out who was actually a fit, and coordinating interviews all competed with the work of running the store.
For Jonathan, the ideal hiring experience wasn't one with more tools. It was one that helped him make the right decision with as little administrative work as possible.
Employer flow
Step 1
Post a job
Post a job → define candidate preferences
We separated the job from the candidate. Employers described the role first, then told Fortuna what a good candidate looked like.
Step 2
Surface the signal
Candidate list → profile → video → interview response
Instead of asking employers to work through a stack of applications, Fortuna surfaced progressively more information about each candidate, from fit and experience to video and interview responses.
Once the job was posted, the challenge became one of information density. An employer shouldn't have to open a dozen full applications just to figure out where to start. The candidate list gave Jonathan an initial signal about fit; opening a candidate revealed their experience and attributes; and the video resume gave him a more personal sense of the person behind the application.
Rather than presenting everything at once, the flow let Jonathan decide how much information he needed before moving to the next step.
Step 3
Interest into action
Contact → interview type → available times → confirmation
Expressing interest shouldn't create another administrative task. Once Jonathan found someone he was interested in, the next source of friction was coordination. Contacting a candidate could easily become another round of messages, phone calls, and back-and-forth scheduling.
Fortuna kept that process inside the same flow. Jonathan could choose an interview type, provide one or more possible times, and confirm, without leaving the candidate experience. Finding the right candidate shouldn't create a new administrative task just to get an interview on the calendar.
Step 4
Keep it moving
Interviews → connect → contact information
Once both sides were interested, Fortuna made the connection explicit and handed the relationship off to the employer and candidate.
The product's job wasn't to manage the entire employment relationship. It was to remove enough friction to get two people who wanted to talk to each other into the same room, or onto the same call.
Under the hood
None of these screens happened in one pass. Behind the final app are rounds of prototypes, a UI kit built from scratch, and a pile of assets made by hand.
The UI kit
One set of colours, type and components shared by both sides, so the job seeker and employer apps felt like the same product.
From prototype to final
Every round of testing fed the next prototype. Here's how the core screens changed on the way to the version we shipped.
Assets
The icons, illustrations and imagery that gave Fortuna its personality.
On the cutting-room floor
fillerA line on the sprint docs: not everything we designed made it in, and some of what got cut taught us the most.
Maybe in the middle
Swipes on Tinder work because attraction is often immediate, but choosing a job is a little different. It means considering the commute, reading responsibilities, weighing tradeoffs, and imagining your future. These are slower cognitive tasks that don't always lend themselves to an immediate yes or no.
Maybe some users wanted to save a job for later, or maybe they wanted to compare it to another listing. Maybe they just wanted a little bit more time to think on it. Maybe we needed a maybe.
fillerWhether Maybe shipped or was cut, and why. (If it shipped, this beat moves back into The swipe.)
Another idea we killed
fillerThe idea, why it was cut, and what it taught us.
One that changed shape
fillerWhat it started as, what it became, and why.
Brand in the worldneeds a title
fillerA line on the brand: what it needed to say to job seekers and to employers, and where it showed up.
Website
App Store screens
Video resume
Promotional shoots and flyers
Closing reflectionneeds a title
fillerOutcomes — any numbers you have — then what you'd do differently, and what you carried into later work.