Nitya Venkataraman · Product Designer

PropertyGuru · Singapore · 2021–23

Featured Agent

A premium listing product that had been sold over the phone for years. I replaced the account-manager booking process with a self-serve product agents could buy themselves, any time of day.

SGD $5M
Revenue reached in 2022, after three flat years
24/7
Booking, replacing office-hours-only phone calls
3–4 days
Queue wait per agent ahead, removed entirely
20+
Agents in post-launch testing panels
Role
Sr. Product Designer
Company
PropertyGuru
Timeframe
2021–2023
Platform
Mobile & desktop web

Featured Agent is a premium product that gives an agent maximum visibility — it highlights their listing among the top seven on the first page of search results, and features their profile on the first page of the search results page.

Open the working prototype →

Background

Featured Agent had seen stagnant revenue for three years, with consistently low fill rates. Despite its potential as a premium listing product, adoption stayed limited — purchases were underwhelming whether agents bought directly or went through a sales representative. Repeat purchase rates were declining, which told us engagement wasn't sustaining after the first booking.

The goals we set:

  • Grow Featured Agent revenue in Singapore to SGD $5M in 2022.
  • Increase visibility, transparency and awareness of the product.
  • Minimise the friction of manual operations.

How booking actually worked

Discovery

Nothing about buying Featured Agent happened in the product. An agent phoned their account manager to ask what was available. The account manager checked, came back, and pushed for a purchase if a slot was open. If nothing was open, the agent joined a queue the account manager tracked by hand — and when a slot freed up, they had 24 hours to confirm by phone before it passed to the next agent in line.

Payment was no better. Featured Agent was a cash-term product, so agents read credit card numbers over the phone or waited for a payment link. They had three working days to complete payment or the slot was released back to the queue. Only once all of that cleared could an agent start prioritising listings within the project by boosting, reposting or auto-reposting.

Flow diagram of the original Featured Agent process, showing an agent phoning an account manager, the manager checking slot availability, a manual queue, and a separate card payment step before the listing goes live.
The original task flow. Every orange step is a human doing something by hand. The agent is never in direct contact with the system they are buying from.

Where it broke down

Six bottlenecks

  • Entry point

    Agents had to call in to purchase at all.

    Buying only possible during office hours

  • Human error

    Account managers tracked new inquiries by hand while also working the waiting queue.

    Missed inquiries, lost slots

  • Payment

    A cash-term product paid by reading a card number down the phone, or via a link the account manager sent.

    3 working days to pay, or lose the slot

  • Transparency

    Slot availability and pricing appeared nowhere on the PropertyGuru website or on Agentnet.

    No way to check without phoning

  • Queue visibility

    Agents in the queue couldn't see how many were ahead of them, or how long they'd wait.

    Indefinite, unexplained waits

  • Time

    Agents waited for every agent ahead of them to confirm and pay before their turn came.

    3–4 working days per agent ahead

Who we were designing for

Research

Two archetypes, and they wanted opposite things. The normal user — middling volume, managing a team — needed the product to be fuss-free and to state its benefit up front. The power user already knew the value and just wanted the phone loop out of the way. A self-serve product had to satisfy both without becoming a configuration exercise.

  • Normal User persona: Lisa Tan, mid 30s, team leader, middle-to-lower percentile. Goals are expanding her client network and streamlining her processes; frustrations are getting noticed among ever-growing competition; must-haves are a fuss-free solution, benefits shown up front, and transparency on what she is paying for.
  • Power User persona: Donovan Wong, mid-to-late 30s, top-earning agent in the upper percentile, with goals, frustrations and must-haves mapped alongside his technology comfort.

Three things research made non-negotiable:

  • Browsing. Agents need to see which Featured Agent slots are available per project, estate and district — and each slot must show its limited availability.
  • Booking. Agents need to choose a slot on the platform and manage their own bookings afterwards.
  • Payment. Agents need to pay in an in-built currency — credits — rather than a card transaction brokered by a person.

The solution

A new Featured Agent screen with a project, estate and district catalogue, reachable from the main navigation under its own link. Agents bypass the call-in system entirely and see for themselves what's available. The card layout gives them something to compare across projects — a far more tactile way to browse than extracting the same information from an account manager over the phone.

The Featured Agent catalogue screen on mobile and desktop, annotated with site header, page header, search bar, tab bar, category header and card grid. Cards show four states: 7 Slots Available with Book Now, Fully Booked with Notify Me, 3 Slots Left with Book Now, and an active booking expiring 21 Sep 2021 with Manage.
The catalogue, on both breakpoints. Four card states carry the whole availability story — open, full, running low, and already yours. A full project offers Notify Me instead of a dead end, which is the queue rebuilt as a product feature rather than a spreadsheet an account manager keeps.
Anatomy of a Featured Agent project card: slot availability at the top, project thumbnail, name, booking cost in Prime Credits, active agents, active listings, and a Book Now call to action.
Solving for transparency. Every question an agent used to phone in to ask is on the card: how many slots are open, what it costs, how many agents and listings are already active in that project. The call to action changes with availability, so the card is honest about whether there's anything to buy.
Booking screen and modal for mobile and desktop, numbered one to eleven: header, project cover image, project name and district, project insights, duration selector, start and end dates, cost in Prime Credits, and the confirm booking action.
Solving for payment. Booking is settled in Prime Credits against a balance the agent already holds, so the card-over-the-phone step and its three-day payment window disappear. Duration is picked up front — 30, 90 or 180 days — and the cost and dates resolve before the agent confirms.
Three booking duration scenarios shown side by side, each with an option to extend the booking before it expires.
Durations, with an option to extend. Extending in place was the mechanism for repeat purchase — the metric that had been declining.
End-to-end mobile user flow for Featured Agent self-serve, from browsing the project catalogue through selecting a slot and duration to confirming the booking with credits.
The mobile flow end to end. Browse, select, set duration, confirm — no phone call anywhere in it.

What testing changed

Post-MVP

After the MVP shipped we ran usability sessions through agent panels, speaking to more than 20 agents. Self-serve was well received across every level of agent — but the consistent ask was for more decision-making information on the card and booking screen, not for changes to the flow itself.

"I would like to see the prices of the other listings. The highest to lowest or even the median value."

Agent, panel session · P1

"I would like to see the total amount of WhatsApp clicks for all the listings in this project. Or how much more leads total."

Agent, panel session · P2

"I would like to see the amount of unsold units in this particular project. If possible also the developer commission here."

Agent, panel session · P3

The through-line was that agents wanted to judge whether a slot was worth buying, not just whether it was available. We added the insights they asked for — times searched, total leads and times shortlisted — to the project insights block in subsequent releases.

The shipped mobile booking screen for Normanton Park, District 5, showing project insights, a 30/90/180-day duration selector, start and end dates, a cost of 24 Prime Credits against a 100-credit balance, and a Book Now button.
The shipped booking screen, with the insights agents asked for.

Outcome

Featured Agent revenue in Singapore reached SGD $5M in 2022 — the target we set, and the first growth after three flat years. The booking process moved from office-hours phone calls and a hand-managed queue to a product agents operate themselves.