Fintech
Designing clarity for portfolio, positions, movements, and financial operations
White-label trading and portfolio platform for stockbrokers. Dense screens where clarity and operations had to coexist.
WombiProduct DesignTwo-sided marketplaceUGC · Argentina
Design the flow that turns a content need into a concrete collaboration.
Wombi connects brands with UGC creators. I shaped campaign, application, evaluation, and selection until intent became collaboration.
- Role
- Co-founder & Product Designer
- Scope
- End-to-end: research, UX/UI, prototyping, functional definition, and user testing
- Team
- Dev, marketing, and customer support
- Timeline
- Jan 2025 - Mar 2025 · Campaign flow redesign
01Context
A brand does not start by looking for a creator. It starts with a content need.
Wombi is a UGC marketplace in Argentina.
Brands need videos for social, ads, or Mercado Libre. Creators produce them.
The problem was not showing a catalog.
A content need is not yet an opportunity a creator can evaluate.
Context is missing: what is expected, how compensation works, which conditions apply.
That context is not symmetric: brands and creators need different information.
Designing the flow meant deciding what appears, for whom, and when.
01Content need
02Campaign
03Opportunity
04Applications
05Selection
06Collaboration
02The two-sided challenge
The same detail that helps a brand can create friction for a creator.
Each side enters the product with a different question. The system had to answer both without penalizing either.
The brand
Wants results and control without becoming a brief expert.
- Define what it wants and what content it expects
- Set conditions and compensation
- Receive relevant applications, not noise
- Compare heterogeneous profiles
- Decide with confidence and move forward
Shared context
The creator
Protects their time: they do not want to apply blindly.
- Discover opportunities that fit
- Quickly understand what is expected
- Know how much and how they will be compensated
- Assess whether they meet the requirements
- Apply and understand what state they are in
Designing a marketplace was not about connecting supply and demand. It was about designing enough context for both sides to decide without asking too much of either.
03User journey
Six stages, three actors, a different risk at each one.
I rebuilt the journey as a collaboration acquisition funnel. Each stage has an actor who decides and a specific way to fail.
| Stage | Actor | Decision | Risk |
|---|---|---|---|
01Create campaign | Brand | Define the need clearly | Ambiguous campaign → irrelevant applications |
02Publish | System | Validate and moderate before opening | Off-policy content slips through late |
03Discover and evaluate | Creator | Assess fit before investing time | Applying without understanding conditions |
04Apply | Creator | Propose price, timeline, and pitch | Weak application → does not help decide |
05Evaluate and select | Brand | Choose one profile among many | Not enough context to compare |
06Start collaboration | Brand + Creator | Confirm and move to production | Confusing transition between stages |
04Turning a need into a campaign
The problem was not asking for more information. It was finding what changed application quality.
Campaign creation is a seven-step wizard.
General info, product, preferences, deliverables, compensation, questions, and summary.
Progressive and grouped by decision, not by form.
Every extra field improves context for the creator, but pushes the brand closer to drop-off.
The answer was not fewer fields, but hierarchy: required what changes matching, optional what refines it.
Compensation
Define how you will compensate the creator for each campaign video
Paid
The creator receives a money payment for their work
Barter
The creator receives your product or service as payment
Paid and barter
The creator receives a money payment and your product or service
$120,000
The creator will see up to $103,200 as payment per video.
- The stepper keeps progress visible: it reduces the feeling of an endless form.
- Compensation is chosen as a model, not as a loose field.
- “The creator will see up to $X” translates price into the metric the brand cares about.
Decisión de diseño
Progress by decision, not by form
Seven steps with a stepper always visible. Each step groups one decision (what I promote, how I compensate, who I prefer) instead of a flat list of inputs.
Decisión de diseño
Required only where it moves the needle
Name, summary, platform, objective, mobility, deliverables, and compensation are required: they bring the right creator. Category, preferences, and questions stay optional.
Decisión de diseño
Conditionals to avoid showing too much
Product shipping only appears if the product is physical. Price fields only if there is monetary compensation. Each branch shows when (and only when) it applies.
05Designing the brief
The brief is the bridge between business intent and creative decision.
In Wombi the brief is not a single text field.
It is built in the wizard and read as a structured piece in the campaign detail.
I offered two levels: creative freedom or a detailed script.
Brands without experience are not blocked; those with criteria can be precise.
Decisión de diseño
Two specification levels instead of a free field
Creative freedom vs. detailed script. It is an explicit brand decision about how much control to give up, and it gives the creator a clear expectation before applying.
Questions that guided the design
- What does the creator need to know to decide and produce well?
- What information creates ambiguity if left open?
- What is required and what can be optional?
- How do we avoid vague briefs without forcing the brand to write too much?
On AI · There is no AI assistance to write the brief. Guided fields, examples, and placeholders provide scaffolding: less fear of the blank page, without delegating the brand intent.
06Making conditions visible
An attractive opportunity stops being relevant if conditions appear too late.
Economic and participation conditions are what decide whether a creator applies. The design answers those questions before the CTA, not after.
CompensationPaid, barter, or mixed? How much do I receive?
ProductWill they send product? Do I need to receive it?
RequirementsDo I meet the age, gender, or location asked?
WorkWhat do I deliver and in what format?
Compensation modes
Paid
Money payment for each video.
Barter
Product or service as payment.
Paid + barter
Money and product combined.
Not every condition exists at once in each campaign: they appear based on compensation type, product, and preferences the brand defined.
07Discovering and evaluating a campaign
The creator needs to assess fit before investing time in applying.
Campaigns
AvailableAppliedRequests
Invited1 VideoSkincarePaid + Barter
Video reviews of our new skincare line
$45,000 - $120,000
3 VideosFoodBarter
Unboxing of our specialty coffee box
$0 (barter only) · includes product barter
- The card shows compensation and video count before anything else.
- There is no text CTA: the whole card is the entry to the detail. Less visual noise.
- “Invited” campaigns rise to the top and are tagged: priority without a separate screen.
The listing uses lean cards: videos, category, compensation, title, and price range.
Just enough to decide whether opening the detail is worth it.
Brand and platforms do not appear on the card.
Hypothesis: prioritize opportunity and compensation over “who”, to evaluate fit by the work.
Decisión de diseño
Lean cards, fast decision
Design hypothesis: by not showing the brand on the card, the creator evaluates the opportunity by the work and compensation. It remains a candidate to validate with listing→detail conversion data.
08Designing the application
Give the brand enough context without making the creator rebuild their profile.
The application asks only for contextual data: price, timeline, revisions, message, and answers to the brand.
Age, location, and portfolio come from the profile.
Portfolio is not chosen per campaign: it reduces friction, at the cost of losing curation per opportunity.
A conscious trade-off in favor of application rate.
Inherited from profile
- Age
- Location (province/city)
- Portfolio (up to 6 videos)
Asked per campaign
- Price per video
- Delivery time (4–10 days)
- Revisions (1–5)
- Message
- Answers to the brand
Decisión de diseño
Reuse the profile, ask only what is contextual
The creator does not rebuild who they are for every campaign. The application focuses on what changes between one opportunity and another (price, timeline, fit) and the rest is inherited from the profile.
Actionable compensation · “Your earnings” appears when quoting: if they deliver in 4 days or less, they receive 100% and save the 15% commission. Compensation becomes actionable, not an abstract number.
09From many applications to one decision
Receiving applications does not solve the problem. The brand still has to decide.
Applications arrive in the campaign “Requests” tab.
The list prioritizes what is comparable: portfolio, name, rating, location, and total price with taxes.
“What you see is what you pay” aligns comparison.
Evaluation happens in a drawer, not on another screen.
The brand reviews, decides, and moves to the next without losing the list.
Creator application
P
Paula P.
CABA · ★ 4.8 (23)
5-day delivery2 revisions
Message
Hi! I am Paula, I have been making skincare content for 3 years. I will approach this brief with an honest unboxing + 7-day use testimonial. I have worked with brands in the category and I take great care with natural light.
Portfolio
▶
▶
Total to pay$142.318
Creator payment$103.200
Service fee$14.376
Subtotal$117.576
VAT [21%]$24.691
Total$142.318
When you approve, you proceed to payment. Once done, chat with the creator is activated. Funds stay protected until you approve the content.
- The list price is the total with taxes: the brand compares apples to apples.
- The drawer lets you evaluate without leaving the list: exploration and decision coexist.
- The “Total to pay” breakdown is collapsible: decision first, detail later.
- Copy anticipates what happens on approve: payment, chat, and protected funds.
Aprendizaje
An honest gap: comparison and filters
There is no side-by-side comparison view. “Most relevant” is an unexplained score. Some filters appear in the UI but do not filter. Real product debt, documented.
10Hierarchy for evaluating creators
The decision happens in reading layers, not all at once.
I ordered drawer information by how an application is actually read: from a quick filter to a decision.
1
Can this profile fit?
Identity, location, rating: the first-seconds filter.
2
What quality and style do they have?
Portfolio in vertical format, as content is consumed.
3
What do they propose?
Message, timeline, revisions, and price: the concrete offer.
4
Do I move forward or not?
Total to pay and what happens next. The action.
11States and expectations
In a marketplace, much of the experience happens while one side waits for the other.
Designing states meant designing the wait: what the creator sees between applying and getting a response, and what the brand sees while work progresses.
I do not claim this “reduces anxiety”: I have no quantitative evidence. It does aim for state and next step to always be legible.
Application sent
The creator cannot apply again; they can edit or delete.
Accepted
The collaboration is in progress. They are notified inside and outside the app.
Rejected
Terminal state, communicated without ambiguity.
12The selection moment
Selecting is not “accepting”. It changes system state and both sides’ expectations.
The button says “Approve”, but the real commitment happens at checkout.
There the brand confirms payment, the collaboration is created, and chat opens.
Separating approve from payment avoids accidental acceptances.
Copy supports trust: “funds stay protected until you approve the content”.
Decisión de diseño
Separate “approve” from payment
Approval opens checkout; payment completes it. Two steps, one clear commitment moment, with escrow explained before paying.
Your content starts with this payment
When you pay, chat with the creator is enabled to coordinate content production.
Available balance$260.000
Total to pay$142.318
Protected paymentsuntil content is delivered
Personalized supportto resolve any issue
Direct communicationwith the content creator
13From application to collaboration
The end of the acquisition funnel was the start of an operational relationship.
With selection, the product changes mode: from comparing candidates to coordinating work with one person.
Before · open campaign
- Open campaign
- Multiple candidates
- Explore and compare
- Decision pending
After · collaboration
- 1-to-1 relationship
- Chat enabled
- Timeline and deliverables
- Shared expectations
Collaboration reframes everything: Messages, Deliverables, Brief, and timeline by state.
If there is product shipping, it adds address and dispatch steps.
Actions shift from accept/reject to coordinate, deliver, and approve.
14Real constraints
The product was designed around operational constraints, not in a vacuum.
I chose four constraints confirmed in the product that forced concrete design decisions.
Physical product shipping
What could happen
Without address and dispatch, the collaboration cannot start in “content in progress”.
Decision
Prior states (address, shipping, product in transit), guide banner, and fallback if the creator does not confirm receipt.
Barter compensation
What could happen
Without money payment, “price per video” does not apply and the proposal becomes confusing.
Decision
Dedicated mode: price is hidden, product value and commission are requested separately. “Barter only” without ambiguity.
Contact data moderation
What could happen
Sharing contact takes the relationship off the marketplace and loses payment protection.
Decision
Block in application and chat with clear copy, plus a preventive banner before writing.
Optional max price and above-cap applications
What could happen
If the brand lowers the cap, applications that exceed it remain.
Decision
They stay visible with an “Above max” badge. On accept, the creator’s original price is charged.
15Designing with moderation and trust
A marketplace must facilitate connection without losing control of what breaks it.
Contact moderation is part of the experience, not a technical detail.
It is anticipated with banners, explained when blocked, and escalates after repeated attempts.
Trust is also designed with language.
At checkout, escrow and direct communication appear right before paying.
Real moderation rules
- Detection of email, phone, social handles, external URLs, and DM requests
- Block with clear copy in application and chat, not silent
- Escalation after repeated attempts (temporary send pause)
- Video platform whitelist so legitimate portfolios are not blocked
“We detected contact details in your message (social media, email, phone). Remove them to send it.”
16Product evolution
The flow grew as new scenarios appeared.
I do not have exact dates for each change. I do show how the flow grew as the marketplace added cases.
01Base
Campaign → application → acceptance → simple collaboration.
02Flexible compensation
Barter and mixed appear: one economic model per modality.
03Physical product
Shipping adds address, dispatch, and receipt states.
04Preferences and reach
Profile filters with reach estimate for the brand.
05Moderation
Contact rules and campaigns that enter review.
Source to reconstruct evolution: components, states, and flows present in the product today. Not a dated timeline.
17Impact
Product scale and design impact.
Scale numbers reflect the marketplace today. Impact metrics are internal estimates from the flow redesign, pending validation with analytics.
Platform scale
1,500+
Creators
Marketplace scale · review periodically
600+
Brands
Marketplace scale · review periodically
500+
Campaigns
Marketplace scale · review periodically
455
Screens designed
Frames audited in Figma (docs/screens)
Design impact · to confirm
~14%
Listing → application conversion
Internal estimate · review
~9
Applications per campaign (median)
Internal estimate · review
~22%
Selection rate (application → collaboration)
Internal estimate · review
-18%
Drop-off in campaign creation
vs. previous wizard · internal estimate
Scale metrics are reviewable periodically. Design impact metrics are internal estimates until confirmed with analytics sources.
Más proyectos
Mass-market products • Own projects
Turning fragmented local information into a scalable discovery product
Scattered local data slowed lodge discovery. I structured filters, trust, and content so the product could scale through organic search too.