Demand was never the problem.
Every lost-and-found platform I looked at failed on the same thing: too few people reporting what they had found.
← All projects
Early-stage startup · Product manager · Team of 8
Kittii
What's yours, found
The live product site: scroll inside. Open in a new tab ↗
Problem statement
Frequent public transit commuters struggle to get lost belongings back, because finders have no easy way or reason to report what they find. Recovery depends on whoever happens to be holding the item, and the owner has no way to reach them.
Kittii is an AI-powered lost-and-found marketplace. Seekers report what they lost, a community of finders called Sherlocks report what they found, and AI matching pairs the two.
Highlights
- MVP
- High-fidelity build, 0 to 1
- 8
- Person team, product lead
- 39%
- Of 54 surveyed never got a lost item back
- 79%
- Want anonymity when returning one
Tools used
Gemini Multimodal Models
Firebase
Google Cloud Functions
Figma
Claude Code
Miro
ClickUp
GitHub
Antigravity
Gamma
What I owned
- Vision & AI Architecture. Defining the multimodal AI matching workflows, Gemini integration, and the core mechanics connecting Seekers and Sherlocks.
- Discovery & Validation. Running a 54-person user survey and interviews across major Canadian hubs, mapping competitor gaps, and establishing baseline market sizing.
- UX & Workflow Design. Architecting frictionless reporting and verification flows to remove user friction and establish trust.
- GTM & Growth Strategy. Designing pre-launch acquisition funnels, viral loop mechanics, and the #FindWithKittii campaign architecture.
As a co-founder, I translated our core vision into a rigorous product roadmap, leading an 8-person cross-functional team spanning engineering, UI/UX design, and marketing to bring Kittii from concept to high-fidelity MVP.
Go-to-market Cracking the Supply Problem: Pre-Launch GTM & #FindWithKittii Contest Mechanics Explore our actual whiteboard sessions mapping out pre-launch acquisition, influencer triggers, and the 200-user growth target. View Whiteboard StrategyHide Whiteboard Strategy
Planning sessions for the go-to-market strategy. The contest flow became #foundatkittii, the pre-launch campaign set the 200-user acquisition target, and the three-phase plan mapped channel assignments across launch week.
Why We Stopped Building and Started with 54 Interviews
Kittii began with friends from my post-graduate programme at Cape Breton University. The founder brought the idea to two of us, and it had already been put in front of Volta, the Atlantic Canada startup house, reaching the second round without getting through.
The feedback was that the idea was promising but the fundamentals were missing. That set the work for the next stage: do the discovery properly before building anything.
The target segment was frequent public transit commuters. For them, recovery today is informal and slow: it depends entirely on whoever happens to be holding the item, and the person who lost it has no way to reach them.
Every competitor ran short of finders
The platform is built for Seekers, but it only works because of Sherlocks.
A marketplace with two sides has a chicken-and-egg problem from day one. You cannot attract Seekers without Sherlocks reporting found items, and Sherlocks will not join a platform nobody has heard of.
Finders had no reason to report.
The database stayed empty and Seekers left disappointed.
Cash is not what moves finders.
61%of finder-motivation responses in our survey were altruism, and 31% were convenience. Only 6% were a cash reward.
Trust and privacy stop them.
Finders cited doubts about legitimacy (24%) and not wanting to reveal their identity (24%) as top reasons for not using an app.
Timing compounds it.
People who lose something need help inside a narrow panic window. Miss it and they have already given up.
So we flipped the model. Lead with Sherlocks, make reporting fast, verified and anonymous by default, and let that supply base pull Seekers in.
The Chicken-and-Egg Trap Behind Lost Items
Competitor insights
Why the incumbents left room
We ran discovery before committing to a build, rather than treating the idea as validated because it sounded good.
Demand is real and constant.
People lose things, and the recovery path is informal and slow.
Institutions, not people.
Almost every established platform was American and sold to institutions: airports, transit authorities, venues.
Organisation-first catalogues.
They catalogue what an organisation has found. None started from the person who lost the item.
No community of finders.
None built a community of finders on the other side, which is the gap Kittii was designed around.
User survey validation
What the survey said
We surveyed 54 people, mostly working professionals aged 25 to 34 in Calgary, Halifax and Toronto.
The losses are worth chasing.
of lost items were worth $100 to $500, and 40% were electronics. Stress averaged 3.2 out of 5.
People want one place to look.
said a single dedicated platform is necessary.
Recovery is improvised.
Most people retraced their steps (41%) or called the venue (28%). 16% did nothing at all.
Searching is the pain.
Too much effort and cluttered, hard-to-search listings tied as the biggest annoyance at 21% each. 47% spent an hour or more looking.
Primary user survey, 54 respondents. Multi-select figures are shares of responses. Read the report ↗
Business model Untested
Kittii intended revenue model
A hybrid B2B2C model: transactions up front, with SaaS underneath. Financial projections and pricing tiers designed for pre-seed monetization.
B2C pricing
Value-based pricing tiers
Platform fees scale with item value, capturing maximum willingness to pay at the moment of reunion.
Tier 1
Basic
$6.99
Keys, clothing, bags, small accessories
Auto-coordinated returns
Up to 4 days for claim processing
Tier 2
Mid-Value
$12.99
Wallets, headphones, sunglasses, small electronics, IDs, pets
Priority matching
Up to 3 days for claim processing
Tier 3
High-Value
$21.99
Phones, laptops, tablets, cameras
Priority + Safe-Zone handoffs
Up to 2 days for claim processing
Tier 4
Premium
$29.99
Jewelry, luxury goods, passports
Full KYC + Chain of Custody
Up to 2 days for claim processing
Priority processing is available for a charge. It reduces the window to 48 hours for Tiers 1 and 2, and 24 hours for Tiers 3 and 4.
B2B SaaS revenue
Partner venues: nominal software maintenance fee
Small Businesses
$99/month
Up to 200 reports
49¢ per report
Mid-scale Businesses
$199/month
Up to 500 reports
39¢ per report
Large-scale Businesses
$580/month
Up to 2,000 reports
29¢ per report
- $30K to $55K
- Year 1 revenue target
- $3K to $5K
- MRR by month 12
Flipping the Model: Leading with Sherlocks
The MVP successfully reached high-fidelity build completion, delivering a dual-sided consumer app for Seekers and Sherlocks, backed by a secure human-in-the-loop admin console to manually verify ownership claims before any item handover.
In the app
The admin console
A Seeker reports what they lost, a Sherlock reports what they found, and AI matches them by match relevancy.
Building for the 0-to-1 MVP
#foundatkittii, a five-day treasure hunt
One question shaped the whole plan: how do we get 100 or more people to actively report found items, in five days, in one city, for almost nothing?
How it worked
- Ten partner shops. Across Halifax's downtown peninsula, each with a branded standee at the payment counter.
- One clue, one QR code. Every standee carried a unique clue tied to one item, linking to the submission form.
- A $50 prize. The first matching report using a shop's clue won a gift card, redeemable only at that shop.
- Paid search plus social. Search intercepted "lost item Halifax" queries at peak intent, while Reddit and Instagram carried the campaign locally.
It turned ten business owners into acquisition channels for $50 each. No sales team, no cold outreach, and the shop's own foot traffic doing the distribution.
Traction
Validated demand and a staged rollout
Validated problem
- 39%
- Never got a lost item back
- 79%
- Want anonymity when returning one
- 62%
- Want one dedicated platform
Primary user survey, 54 respondents. Read the report ↗
Product and planned rollout
Live prototypekittii.ca ↗- Alpha
About 30 users from a closed circle
- Closed beta
A closed event for up to 250 people
Raw whiteboard sessions mapping pre-launch acquisition, influencer triggers, and the 200-user growth target.
| Metric | Target |
|---|---|
| Total submissions | 200 to 250 |
| Submissions per side | 100 each |
| Recovery rate | 30% |
| Instagram followers | 500 |
A clear pass or fail, on purpose
The beta was capped at 200 to 250 submissions. A deliberate, controlled cohort rather than a push for scale.
Each number was set so the beta would give a clear yes or no, rather than numbers that merely looked encouraging. A 30 per cent recovery rate is the one that matters, because it is the only measure of whether the product delivers end-to-end value.
This plan was never executed. The team wound down before launch, so the campaign never ran and none of these targets were tested.
Reflection
Two things I would change, both visible only in hindsight.
- We set a revenue target before we had a reliable way to bring in either side. A product with two sides is worth nothing until both sides are there at the same time, and the plan treated finders and seekers as two problems to solve one after the other, instead of one problem. The GTM eventually corrected for this, but it came late.
- I would have tested the matching quality much earlier, against a small real corpus of lost and found reports. AI matching was the differentiator on paper, and it was the part with the least evidence behind it when the build started.
Where it stands. The build reached MVP and stopped there. The go-to-market plan was written but never run, and the revenue target was never tested. The product site stays up as a preview, though the product itself never launched.
Back to the work
Four projects: Lumi Cards, Side Quest, HelloFresh and Kittii.