# Reinventing home search
Liv.rent · 2018 · Product Designer · Case study
Source : https://albancarmet.com/case-study/designing-a-housing-platform
A two-sided rental platform, from listing search to lease signature.
In one line : The whole rental journey in one app
Team : PM, EM, 6 engineers and me · Duration : 6 months · AI : None

## At a glance

- Problem : Liv’s internal MVP, not yet released, only let renters search, chat and book viewings — with heavy processes and little to build trust, for renters as for landlords.
- What I did : As the only designer, I took the app from its internal MVP to public launch: I redesigned the first features and designed the new ones, five major features on both sides of the platform, from listing creation to lease signature.
- Result : A rental journey covered end to end, up to paying rent, with five major features designed or redesigned.
- Limit : No budget for usability tests: every decision rests on agent feedback and judgment, and the component library I was building out struggled to keep pace.

## The company

Liv.rent is a rental platform born in Canada that brings renters and landlords together: verified listings, tenant screening, lease and rent payment in one app. It now describes itself as “the safest rental platform in North America”. When I joined, it was a first iOS version, internal and not yet released, built in Vancouver by the Machobear studio.
- Sector : Proptech · rental
- For : Renters, landlords and agents
- My time there : 2017 – 2018 · Product Designer

## The product : Liv

The app covers the whole rental journey, on both sides: what a landlord publishes is what a renter searches, visits and ends up signing.
- Publish : Landlords and agents create a listing: photos, rules, lease terms.
- Search and visit : Renters browse listings, chat with landlords and book viewings.
- Sign and pay : The lease is created and signed in the app, where rent is then paid.

## Process

No usability tests: there was no budget for them. And no AI: this was 2018.
- Start from research : My earlier research on rental search, for my own app Elly: scams, and hours lost. (without AI)
- Listen to agents : Real estate agents used the app early and passed on what renters kept asking. (without AI)
- Audit the flows : Mapping every step for the heaviest case: a property manager entering every rule and term. (without AI)
- Design : Five major features, as the only designer, on both sides of the platform. (without AI)
- Ship : On iOS, up to lease signature and rent payment. (without AI)

## The problem — A rental app, not yet released, that only let you search, chat and book viewings

Renters lost hours to outdated ads and feared scams. Landlords wanted solvent, respectful tenants. The internal first version left heavy processes on both sides, with little to build trust.

## The approach — Two pillars and no test budget

The first version had set two pillars: make the search effective, and build trust on both sides. Every feature I designed extended one of them, with agents as the main source of feedback.

## The trade-offs — Three decisions made on judgment

With no tests to settle them, these choices rest on reasoning and on what agents reported. Here is what each one gave up.

## What shipped — From the listing to the signed lease

Three of the features that took the app beyond search: a listing that reads at a glance, a score that builds trust, and a lease signed in the app.

## Constraints

- One designer, no testing budget : Feedback came from the product manager and from real estate agents. Every decision was made without formal user validation.
- Requirements that kept moving : A startup with limited funding, where new features arrived while others were still being designed.
- A component library outpaced by features : A small component library existed when I arrived, with no documentation or usage rules. I built out most of it, adding a large number of components, but every redesigned flow made others obsolete, and keeping it up to date at the pace of features wasn’t possible.

## Signals and research

No usability tests on this project — there was no budget for them. Here is what the decisions rested on instead.
- My earlier research · Elly — Scams, and a search that eats time : Before Liv, I researched room and roommate search for my own app, Elly. The same two pains held for renters: scams on listing sites, and hours lost to outdated ads and follow-ups.
- Real estate agents — What renters kept asking : Agents used the app early and passed on complaints. Listing the utilities not included in the rent came straight from them.
- Flow audit — Mapping the worst case first : Before adding anything to listing creation, I mapped every step for the heaviest case: a property manager entering every rule, term and agent.
- Product direction — Two pillars set by the MVP : Machobear had built the MVP on two pillars: make the search effective, and build trust on both sides. The features I designed extended both.

## Decisions and trade-offs

- A vertical photo list rather than a horizontal carousel. Rejected : Thumbnails under a full-size photo — A strip of thumbnails under a full-size preview, swiped like the listing’s own carousel. Delete sat over the full-size photo. Kept : One column of large photos — Large photos in a single column, each with a delete icon and a drag handle always visible. The top photo is the cover. Why : Scrolling vertically was more convenient than swiping, and the delete icon got lost over a full-size photo. What it cost : The full-size preview, and consistency with the listing’s own carousel.
- Button states rather than error messages. Rejected : One error per missing field — A separate error message for each missing piece of information, on every step of listing creation. Kept : An inactive Next button — Next stays inactive while something is missing, under a progress bar that shows how far along the user is. Why : Avoid a flood of error messages in a listing flow that was already heavy, and got heavier with every new feature. What it cost : Precision: the button says something is missing, not what.
- Condensed lease details rather than everything shown. Rejected : Every term on the listing — The previous listing page, with the price over the photo and every detail in one long scroll — and many new terms still to add. Kept : Key facts first, details on demand — A top block with price, rental type and key features; below, a lease section with secondary terms behind “view details”. Why : Let renters scan the key lease terms without pushing amenities to the bottom of the page. What it cost : Seeing every term at a glance: secondary terms now take an extra tap.

## What shipped

- A listing that reads at a glance (The rent price legible again, and key facts before details.) : problem : The rent price sat over the photo behind a dark overlay, with poor contrast — and the listing had many more details to hold. Solution : I kept the immersive photo, replaced the overlay with a subtle gradient, and split the page: key facts first, details below.
- The Liv score (A trust label landlords can read, and renters can improve.) : problem : Landlords needed to judge a renter’s reliability, and renters had no way to prove theirs. Solution : A score in points, based on credit history and profession, in four categories from Poor to Excellent — each with its own color and a recommendation.
- Leases created and signed in the app (From the accepted application to a lease signed by both sides.) : problem : Liv had no way to create, send or sign a lease: that part of the journey happened outside the app. Solution : Lease creation reuses the application and listing data, in tabs with coach marks. Signing works by typing or drawing, with a notification at each milestone.

## Results

The whole rental journey in one app, designed without a single usability test.
(Liv.rent · iOS · 2018)
- 5 : Major features designed or redesigned, on both sides of the platform
- End to end : From listing search to lease signature and rent payment
- 0 : Usability tests: every decision was made without formal validation
What worked:
- Renter profiles save time on both sides: no reintroducing yourself in every chat, and applications come prefilled.
- The whole rental journey lives in one app, for renters and landlords alike.
- Working with illustrators gave administrative flows a human touch.
What didn't work:
- Visual inconsistencies crept in as features shipped faster than the component library.
- Without tests, some components were never challenged — the listing activity section could have been simpler.

## What I took from it

A design system needs dedicated time. Maintaining a component library with no documentation or usage rules alongside feature work was a losing battle. Today I’d build a real design system and set time aside for it, as a deliverable in its own right rather than a parallel task.