back
Trackmob
Recurring donations didn't work in Brazil. We built the infrastructure that made them work.
Role
Co-founder
Lead Product Designer
Lead Product Manager
Team
+2 designers
+3 product managers
40+ other employees
Period
2012 to 2020
About the company
Trackmob built fundraising technology for the Brazilian nonprofit sector: a donor CRM, a recurring payment gateway, and lifecycle tools covering acquisition through retention. Bootstrapped and grown organically.
—
Tags
0→1 Products
Design Systems
Multi-stakeholder Design
Operational Design
Platform Design
Product Design
Product Leadership
Product Management
User Research
Visual Identity
I co-founded Trackmob and led product and design for eight years, taking it from a field-team app to the payments and CRM backbone that 400+ nonprofits ran on, including Greenpeace, MSF, and WWF. The platform processed R$650M+ in donations.
Logo created for the company, which ended up becoming our lasting identity. We started out as an app for managing field teams.
The problem
In the early 2010s, most Brazilian nonprofits managed donors in spreadsheets. Some had less than that. Organizations couldn't plan beyond a few months, and some closed after losing a single major donor.
But the deeper problem was money. Reliable recurring donation billing simply didn't exist in the country.
The real blocker: banking
Monthly charges meant exchanging CNAB files, a bank format from the 1980s, fragile and badly documented. Banks often blocked the transactions automatically, reading donation patterns as fraud.
Small organizations couldn't access recurring billing at all. Large ones kept teams of three to five people doing nothing but managing charges by hand.
Four groups felt this differently:
Fundraising teams:
Days lost to manual bank files and failed charges.
Leadership:
No way to forecast next month's revenue, so no way to plan or hire.
Field fundraisers:
Banking data collected on paper, in the rain, with no validation.
Donnors:
Gave money, then heard nothing, in a country where nonprofit scandals gave them real reason to distrust.
One problem sat underneath all four: no reliable, integrated system.
How I got to the product
I spent months in the operation before I trusted myself to design for it. I ran interviews with fundraising teams, shadowed how orders and charges actually moved, and studied how the most sophisticated NGOs worked, the ones that had solved pieces of this abroad.
Two things came out of that. First, the sector didn't need a better version of what existed, it needed infrastructure that didn't exist at all. Second, the deepest pain was in the money: the recurring charge that failed, the donor lost between signup and the first debit. That reframing set the order of everything I built next. Fix the plumbing first, then the experience on top of it.
What I built
I led product and design across the whole thing. Four moves mattered most.
The pivot
In 2012 we were building a field-maintenance work-order tool. Then Greenpeace asked us to digitize face-to-face fundraising, the street sign-up process that was still entirely paper. We adapted our field tool into an app for it. That's when we saw the real market: an entire sector running on tools from another era.
F2F, before and after:
Before:
Paper forms
Handwritten bank data
7+ day cycle
Donors lost in the wait
After:
Digital capture
Instant validation
Charge started in under 10 minutes
01.
Basic interface built using the Material UI framework available at the time. It could be customized for each client.
02.
Pages from the P2P campaign created for APAE, an institution that supports children with special needs. In order, we have the P2P registration page followed by the shareable donation page.
The CRM
New clients brought a bigger problem into view. Most organizations had no structured control over their donors, and the sales CRMs a few large ones used, Salesforce and Dynamics, were built for closing deals, not for relationships that last years. That was the core insight: a donor isn't a lead you close, it's a relationship you keep alive. I built the CRM around that: unified donor history, behavior-based segmentation, real-time base health. But it couldn't deliver real value until recurring billing worked. Which is what pushed us to the gateway.
Direct mail feature for new donor acquisition.
The gateway
The hard call. Recurring billing didn't exist, so we built it ourselves. Six months of bank negotiations and an intelligent retry engine that decided timing, number of attempts, and messaging. The first version bit us. One error locked R$200,000 for three days. Another sent duplicate charges. We stabilized it slowly.
Billing success rate:
~70% to ~92%
Small nonprofits ran recurring donations for the first time. Large ones cut their charge teams in half.
01.
Central feature across all products that manages lead and donor data, acting as an integration hub between our tools and external platforms
02.
Tool for donation collection, billing, and retry attempts.
Refactoring under pressure
By 2019 our monolith was the ceiling. Pix and LGPD were coming and we couldn't move fast enough to meet them. Refactoring to microservices while keeping operations live risked freezing the company. We did it anyway. It took eight months and it wasn't clean: one bad deploy froze billing for four hours on end-of-month processing day, affecting R$1.2M in transactions, and the team reprocessed failures by hand all night. That mistake was on me. But we finished without stopping operations or losing a major client, and we integrated Pix in three weeks.
We kept both architectures running simultaneously
New clients went directly into the new architecture. Old clients were migrated progressively, with the old system kept as fallback for at least a month.
Beyond product, being a co-founder meant covering whatever the company needed before it could hire for it. In the early years I structured not just product and design, but development workflow, support, CS, and even HR, handing each off as the company grew past 40 people. I built and led a product team of 12. Remote-first, documentation-heavy throughout.
Results
On average, organizations grew new donation revenue by 45% after adopting our products. We didn't run a controlled test to isolate that. But across most cases, the thing that changed between before and after was our products entering the process.
R$650M+
processed in donations
400+
organizations served (Greenpeace, MSF, WWF, UNHCR)
~70% to ~92%
billing success rate
First-ever
recurring billing access for small nonprofits in Brazil
What I learned
The architecture that takes you from zero to one rarely takes you from one to ten. For years I optimized features. The decision that actually defined whether we would scale or stall was structural, and it carried the highest risk and the least visible payoff. Knowing when to tear down what already works, and having the nerve to do it while the business runs on top of it, is its own skill. I learned it the expensive way.
I sold my stake in 2023, after eight years.
