Redesigning deposit architecture

My Role

Content design, IA, naming, flow logic, and content hierarchy.

Team

Data Science for problem framing, PM, Product Designer, and one Engineer.

Methods

Funnel analysis, competitive benchmarking, IA mapping, and A/B testing.

Problem

77% of users never made a deposit, and without one, the product is empty

Bitso is a crypto exchange focused on Latin America. Deposit is the only way to bring money into the platform. Without a balance, users can't buy, convert, or trade. It's the single most critical action in the entire product.

Data Science flagged a severe drop-off in the deposit funnel. Of 718,660 registered users, only 42,206 completed a deposit: a conversion rate of around 23%. The biggest drop happened at the very first step: users were tapping the deposit button and abandoning immediately.

Deposit funnel chart from registered users to deposit tap, completed deposit, and executed trade — The funnel made the product problem visible: the first deposit step was where intent disappeared.

My role

I owned the content and IA decisions, and identified the root cause

I Led

Root cause diagnosis: the language mismatch
All three IA grouping explorations
Naming and labelling decisions throughout
Content hierarchy in the final flow
Rationale for choosing option 3

Team Owned

Funnel data and problem framing
UI design and component decisions
A/B test setup and analysis
Technical feasibility of ramp options

The insight

Marketing said "buy crypto." The product said "deposit". Users got lost in between

I started by asking a question the funnel data couldn't answer: what mindset do users arrive with? Looking at Bitso's marketing (ads, app store listings, social media) the message was consistently "buy crypto." But inside the product, there was no "buy" button. There was only "deposit."

Users who wanted to buy were hitting a deposit flow and not recognizing it as the path forward. They weren't dropping off because the flow was broken, they were dropping off because the language told them they were in the wrong place.

The secondary problem I found
Even users who understood "deposit" were immediately confronted with a list of 42 currencies, an overwhelming choice architecture that forced them to know what they wanted before the product had explained their options. The first step asked for expert-level knowledge from brand new users.

Comparison between Bitso buy crypto marketing and product deposit entry points — The acquisition promise was "buy crypto," but the product entry point was framed as "deposit."

Original deposit flow with a long undifferentiated list of currencies — What users saw when opening the app.

IA exploration

Three ways to group 42 options. Only one scaled

The core content design challenge was: how do you reduce 42 options to something a new user can navigate in seconds? I explored three grouping models, each with different logic.

Option 1 · Fiat vs crypto

Group by currency type: deposit local currency vs. deposit cryptocurrency. Clean and logical, but assumes users know which type they have.

Option 2 · Your accounts vs others

Group by ownership: deposit from your own accounts vs. receive from someone else. More user-centric but duplicates Bitso Transfer in both groups.

Option 3 · Banks vs wallets vs Bitso Transfer

Group by method: how they're actually moving money. Maps to how users think about their financial tools, not how Bitso categorizes assets.

Chosen

IA diagram: fiat vs crypto grouping — Option 1: fiat vs crypto

IA diagram: your accounts vs others — Option 2: your accounts vs others

IA diagram: banks, wallets, and Bitso transfer — Option 3: banks vs wallets vs Bitso Transfer (chosen)

The decision

Option 3: It scales, avoids duplication, and removes a currency selection step

Why not option 1?
Fiat vs crypto requires users to identify their asset type first. New users often don't think in those terms. Someone who wants to "add money from their bank" shouldn't need to know whether that's "fiat" to find the right path.

Why not option 2?
Ownership grouping duplicates Bitso Transfer and adds cognitive load. Users who encounter the same option twice assume they've made a wrong turn.

Why option 3?
Method-first grouping maps to how users already think about money movement. Everyone knows the difference between a bank, a crypto wallet, and another user. This model requires zero financial literacy to navigate and scales cleanly as new fiat ramps are added.

Before and after comparison of the original deposit list and the new grouped deposit flow — The redesigned flow changed the first decision from "which asset?" to "how are you moving money?"

Results

+15pp deposit conversion in one week

The A/B test ran against the original flow. Users who saw the new grouped architecture converted at 56.5% from deposit tap to completed deposit — compared to 40.1% for the control. That's a 15-percentage-point lift, in one week, with no engineering changes to the underlying deposit mechanics.

56.5%conversion, new architecture

40.1%conversion, control

+15pplift in one week

A/B test results comparing control and experiment deposit conversion rates — The grouped architecture improved conversion from deposit tap to completed deposit without changing the underlying mechanics.

Learnings

What I'd do differently, and what's next

What I'd do differently

Run a quick usability test on the three grouping options before going straight to A/B. We had strong rationale for option 3, but a 30-minute session with 5 users would have validated the naming before shipping, and might have caught the "deposit and receive" label ambiguity earlier.

What the next experiment would be

Address the upstream problem: the "buy crypto" language mismatch. The architecture fix improved the flow, but users who arrive expecting to "buy" still encounter a "deposit" entry point. A buy-first flow, even as a thin layer over the same backend, is the remaining gap.