NoBroker — The 20-Point Gap

Funnel Conversion Real estate · Bangalore

I used NoBroker as a real user
for one afternoon. Same wall. Two states. A 20-point conversion gap.

₹4.76L

estimated monthly recovery
from one trigger change

The signup wall appeared at two completely different moments, sometimes mid-scroll, before a user had found any listing. Other times only after clicking "Get Owner Details" on something they clearly wanted.

Same wall. Same copy. Same design. The only variable was intent, and it produced a 20-point difference in signup completion.

11,000 users were abandoning every month in 12.5 seconds, not because of the wall, but because of when it appeared. Moving one trigger recovers an estimated ₹4.76L/month based on industry benchmarks.

No internal data. No team briefing. One session, one observation, one defensible fix.

Full write-up on Notion Data on GitHub

↓ see the full analysis

2.07%

paid conversion
from 50,000 users

12.5s

average time before
11K users abandoned forever

reactivation rate —
every exit is permanent

₹4.76L

recoverable monthly
revenue with one fix

01 · The Problem

NoBroker promises
zero friction.
It delivers the opposite.

NoBroker is India's largest broker-free property platform. Its core promise: connect renters directly with owners, zero brokerage. The platform runs on a freemium model, 9 free owner contacts, then a paywall from ₹1,099 to ₹7,499.

Out of 50,000 Bangalore rental users per month, only 1,033 paid, a 2.07% end-to-end conversion rate. But the headline number isn't the real story. The mechanism of loss is.

11,000 users started signing up and abandoned in 12.5 seconds on average. Too fast to reach OTP. They saw the phone number field, said no, and left. Not one ever came back.

Funnel — 50,000 users → 1,033 paid

Landed

50,000

Searched

26,100

−47.8%

Hit signup wall

24,200

—

Signup complete

14,000

−42.1%

Contact unlocked

11,200

−20%

Paywall seen

8,120

—

Paid ✓

1,033

−87.3%

12.5s

avg abandonment — too fast to reach OTP

of signup wall abandoners ever returned

02 · The Finding

Same wall.
Wrong moment.
20-point gap.

NoBroker's signup wall isn't broken. It's mistimed. When it fires mid-browse, before a user has seen a single listing they want, it reads as a toll gate. The user has nothing yet. They leave. When it fires post-intent, after they've tried to contact an owner, the ask feels fair. There's something concrete to unlock.

Path A — Mid-browse 60% of users

50%

signup
completion

447

paid
conversions

Wall fires mid-scroll, before the user has found a listing they care about. Phone-only ask at the worst possible moment — asking for something irreversible when the user has nothing yet.

Path B — Intent-driven 40% of users

69.8%

signup
completion

586

paid
conversions

Wall fires after clicking "Get Owner Details" on a specific listing. Same phone ask, but the user knows exactly what they want. The wall is now a door, not a toll gate.

03 · The Fix

One trigger.
₹4.76L
recovered.

Move the signup trigger from mid-browse to post-intent

Sprint 1–3 · Medium effort · No design overhaul

Effort

Medium

Don't ask users to register before they have a reason to. Let them browse. Let them shortlist. When they try to contact an owner or save a listing, that's the moment. They already want something. The wall is now a door.

Offer phone or email at signup. Users who pick email get asked for phone at contact unlock, at that point they're invested. No new screens. One trigger point moved. That's the entire build.

30% recovery of 11,000 monthly abandonments at the mid-range conversion benchmark recovers ₹4.76L/month. Every month this doesn't ship is another 11,000 users permanently gone, at ₹150–₹400 CAC each.

Monthly recovery

₹4.76L

Effort

Sprint 1–3

Risk

Low — one trigger moved

Design change

None required

A note on the numbers

This analysis used no internal NoBroker data. The funnel figures, drop-off rates, session durations, conversion gaps, are modelled using publicly available SaaS and proptech benchmarks. The ₹4.76L monthly recovery estimate is directional, not precise.

What isn't modelled: the two-path observation. That came from using the product as a real user. The numbers show what it might be worth. The session showed that it exists.