171
40%
precision
QUERY: “check-in inventory”
Narrow domain language surfaced deposit disputes, inventory failures and documentation tactics.
Product strategy case study · 2026
Deposit-Proof
A renter is holding ₹4.2 lakh of exposure and a phone video she’ll lose in ten months. I spent a day finding out whether that’s a product.
CaptureRoom-by-room evidence
ShareNeutral landlord link
RecordViewed acknowledgment
- Product strategy, research, interaction design, UX writing
- 1 day
- Concept, unbuilt, and here’s why.
01 / The verdict
The answer belongs before the process.
0.43×
LTV over CAC on paid acquisition, against a 3× bar
40%
precision on the research scrape that worked
~0%
precision on the one that didn’t
The concept is structurally sound.
The standalone consumer model is commercially unviable.
The work that proved it cost a day rather than a funding round.
02 / The pivot
The brief arrived as a decluttering app.
A worthwhile job needed to survive three tests. Decluttering survived none of them.
01
Hard deadline
Possession and move-out create fixed moments.
02
Counterparty
The landlord must accept the shared record.
03
Money at risk
₹1.5–3L can depend on the evidence.
| Opportunity | Hard deadline | Counterparty | ₹ attached to failure |
|---|---|---|---|
| Decluttering | No | No | No |
| Spatial planning | No | No | No |
| Condition evidence | Yes | Yes | Yes |
“A 24–34 year old salaried renter in Bengaluru, on possession day and again on vacating day, needs to prove the condition of a flat to a landlord holding ₹1.5–3L of their money, and currently solves it by shooting an unstructured phone video and hoping WhatsApp still has it ten months later.”
03 / Research
Including the part that failed.
Two scrapes of renter and landlord forums across India, the UK and France. Failure was useful data, not an appendix.
466
~0%
precision
QUERY: “unfair landlord”
Broad emotional queries fell back to popular content and returned relationship drama.
The method finding: broad emotional keyword queries don’t retrieve; narrow jargon does: “check-in inventory”, “état des lieux”, “security deposit deduction”.
04 / Evidence
What the evidence said.
01
The missing check-in inventory decides every case.
“A UK claim went from £400 to £1,835 the moment the tenant refused to settle privately. A landlord, unprompted and in caps, said always pay for check-in reports.”
The decisive artifact isn’t the photo. It’s the counterparty’s acknowledgement of the photo at a time they can’t dispute it.
02
Deductions inflate on escalation and deflate on documentation.
“One tenant cut a £600 claim to £120 just by threatening dispute.”
The output is a negotiating instrument, not an archive.
03
Metadata integrity is already folk knowledge.
“Tenants cite ‘date/time/GPS metadata intact.’ WhatsApp strips metadata.”
Stamps are visible on the artifact, not buried in EXIF.
04
The category’s branded apps rate 3.1–3.4 stars.
“Paywall complaints recur; two products are dead while roundups still recommend them.”
Users install, hit a paywall before value, leave. Full value before payment.
05 / Competitive gap
The empty cell is the strategy.
| Product | Who it’s for | Unit | Price | Sign-off | India |
|---|---|---|---|---|---|
| zInspector | Property teams | Portfolio | Paid | Yes | No |
| RentCheck | Property managers | Portfolio | Paid | Yes | No |
| HappyCo | Operators | Portfolio | Paid | Yes | No |
| NoBroker | Renters + owners | Transaction | Mixed | No | Yes |
| Google Photos | Anyone | Media library | Free | No | Yes |
| Anyone | Message thread | Free | Viewed only | Yes | |
| Tenant-first app | Tenant | 1 home | Not offered | Not offered | Not offered |
The insight
The only condition record your landlord has already agreed to.
Everyone can take photos. Nobody can get the other side on record without looking hostile. Solving that is the product.
06 / Interactive prototype
The product is one low-friction tap.
Play both sides of the exchange. The record only becomes useful when the landlord sees it and feels safe saying so.
The USP: not better photos. A condition record the other side has already acknowledged.
Deposit-Proof
YOUR RECORDYour move-in record is complete.
Flat 4B · Move-in
Indiranagar, Bengaluru
Living
8 photos
Kitchen
6 photos
Bedroom
9 photos
23 Sep 2026 · 09:14 IST
Location stamped on every photo
Hi Rajiv, moved in today, thanks again. I’ve put together a quick photo record of the flat as it is now, so we both have it on file for later.
1 / 3
07 / Design decisions
Every useful choice creates a liability.
1
One-tap “I’ve seen this record” that admits nothing.
- Far weaker than a signature.
- A signature request gets ignored and a viewed-receipt doesn’t.
2
Never mention deposits, damage or disputes on the landlord page.
- The product under-sells itself to half its users.
- She won’t send something adversarial while she still needs the geyser fixed.
3
Move-in capture free forever; payment only at move-out.
- A ten-month gap between value and revenue. That later breaks the economics.
- Paywall-before-value is what killed the incumbents.
4
Append-only captures, with no edit or delete.
- She can’t fix a bad photo.
- A record she could have edited is a record he can dismiss.
08 / UX writing
The copy is the interface.
In a product whose success metric is someone else’s tap, words carry the interaction.
“Confirms you’ve viewed it. Nothing else.”
Pre-empts the exact hesitation that would kill the tap.
“Hi [Name], moved in today, thanks again. I’ve put together a quick photo record of the flat as it is now, so we both have it on file for later.”
Reads as housekeeping, not a claim. Not a trick; it’s accurate, and it’s why he taps.
“Now get him to look at it. A record he’s seen is worth roughly ten times a record he hasn’t.”
Tells her the truth about where the value is.
09 / Future opportunities
The product survives. The route to market has to change.
The failed consumer economics are not the closing argument. They are constraints for the next three business-model tests.
Paid acquisition cannot carry this product.
Only 4% of installs become paying customers. CAC is ₹1,500 against ₹650 LTV, and even a ₹1,499 price reaches just 1.3×.
4%
₹1.5K
₹650
PATH 01
Make sharing the acquisition loop
Every useful record naturally touches a tenant, flatmate and landlord.
- If recipients create their next record from the shared link, distribution can approach zero marginal cost.
- Acknowledgment may create trust without creating enough curiosity to start a new record.
PATH 02
Flip the payer to small landlords
A landlord with several units has repeat usage and a clearer budget than one tenant.
- A lightweight portfolio view could turn occasional tenant utility into recurring landlord value.
- Landlord-first positioning could weaken the renter trust that makes acknowledgment possible.
PATH 03
Borrow trusted distribution
Brokers and relocation firms already sit inside possession-day workflows.
- Bundling the record into handover can remove paid CAC and make completion part of the service.
- Partners may treat evidence capture as extra operational work without a clear commercial benefit.
No path earns a full build. Each starts with the smallest test that could disprove it, and investment follows evidence.
10 / Kill criteria
Written before the data, so they couldn’t be rationalised afterward.
- 1
Acknowledgement below 20% of 100 sent records.
- 2
Move-out return below 10%.
- 3
Fewer than 3 of 8 interviewees having lost material deposit money.
- 4
LTV/CAC below 3× at realistic CAC.
11 / What I’d do next
Test the riskiest tap, not the whole product.
The full product takes six weeks to build. Its riskiest assumption takes an afternoon.
STEP 01
Build the landlord-facing page alone
No renter app, no capture flow, no payments. Just the page he lands on.
STEP 02
Seed it with manually shot photos
One flat, captured by hand, room by room, with visible stamps.
STEP 03
Send it to twenty real landlords
Cold links, no audience, no spend, nothing to maintain.
4
taps of landlord attention
Under four taps, kill criterion one fires. Twenty links settle it with data instead of opinion.
CRITERION 01 · Acknowledgement below 20% of 100 sent records
This test
1 afternoon
Full build
6 weeks
Same evidence, at a fraction of the cost.