Mortgage matrix — the five screens of the main flow

Mortgage matrix — pre-approved limits

Product Designer · Domclick · Mar 2022 – Jan 2023

Choosing a mortgage was overwhelming: too many programs, terms that felt hidden. I redesigned the flow around one idea: the bank pre-approves a maximum limit, and the user chooses freely within it.

Problem

46.2% mortgage refusals, ~half from wrong-program applications

My role

✦ End-to-end product design✦ 3 iterations✦ 15 usability tests

Impact

✦ MAU 10×✦ First pre-approved limits product in Russia

Timeline

Mar 2022 – Jan 2023

Skills

✦ Product thinking✦ User research✦ Design

Team

✦ 3 designers✦ Product Owner✦ Design lead✦ 5 developers✦ Illustrator

Product & Problem

Problem

Choosing a mortgage is overwhelming: too many programs, legal jargon, conditions that feel hidden. Between Q4 2020 and Q4 2022 the bank's mortgage refusals rose from 31.5% to 46.2%, and about half came from people applying to the wrong program.

The core issue: users can't tell which program fits them, and they suspect the bank is hiding the real terms.

The design bet: instead of asking users to navigate the complexity, the bank takes on the risk of choosing terms: it pre-approves a maximum limit, and the user acts freely within it. Fewer wrong applications → fewer refusals.

4th quarter of 202031.5%

4th quarter of 202136.6%

4th quarter of 202246.2%

Mortgage refusals hit a historic high: half from wrong-program applications Source

My Role & Contribution

As the designer on the mortgage-calculator team, the first step in the approval flow, I coordinated designers across the other teams to keep our decisions aligned. I also took on the user research myself, because I wanted to understand our users first-hand.

Key Design Decisions

I ran three iterations with 15 usability tests, reshaping the flow around what each round of testing revealed.

Exploration flow — how the mortgage journey was reshaped

Iteration 1 — establish the value

1st iteration — mockups of the pre-approved limit flow

The first mockups made the core value legible: users understood their maximum available amount, the program conditions and what onboarding asked of them. But testing exposed two blockers: people still couldn't choose a program confidently (so they wouldn't submit), and on the very first screen they didn't understand where they'd landed or what to do. Clarity of value wasn't translating into clarity of action.

Iteration 2 — simplify the choice

2nd iteration — simplified program selection

So I focused on making program selection clearer and cutting the number of choices upfront. This surfaced a sharper set of problems: too many buttons after onboarding, a compare function that wasn't obvious and confusion about the FROM–TO limit the bank had approved. Still too heavy. But one signal was decisive: users loved the underlying idea that they're approved for a maximum and can do anything within it. That validated the concept even though the execution had to get simpler, so I dropped the multi-choice, compare-heavy layout and carried the idea into a third iteration.

Final solution — a single, sequential path

I rebuilt the flow to move one clear step at a time: the maximum amount always visible, conditions shown upfront, onboarding that explains what's coming next. In testing, users finally understood what to do at each step while keeping the freedom of the pre-approved limit. The compare-everything model was gone; the "approved maximum, act within it" idea stayed.

Main user path — the 15 steps from entry point to property chosen

We wanted to give the official legal terms of mortgage programs some kind of identity, so we ordered a set of icons from a 3D illustrator.

Sixteen 3D icons commissioned for the mortgage programmes

Outcome

Result

After launch, monthly active users of the flow grew 10×, and it became the first pre-approved-limits mortgage product on the Russian market. Conversion to a completed deal held steady even as the audience expanded: a much larger top of funnel didn't dilute the quality of applications.

MAU: 10x

Conversion to a completed deal: held steady

A note on data: I moved on from the product shortly after launch, so MAU is the metric I can speak to firsthand.

Reflection

The real unlock here was alignment, not interface: the hard part wasn't the UI but getting the bank comfortable taking on the risk of choosing terms for the user and planning that client path carefully before moving into high-fidelity.

If I did this again, I'd sharpen two things. I'd define success metrics from day one: we grew MAU 10×, but I'd instrument a downstream metric (conversion to an approved deal) from the start, to show we were driving better-qualified applications, not just more of them. And I'd prototype less early on: a couple of the flows I built and tested could have been validated with simpler wireframes first.