Search tells you what's missing. GMV tells you what's working. Every marketplace that scaled ran the loop between these two.
This isn't a lend-it invention. Airbnb, Uber, DoorDash, Etsy, and Turo converged on the same two-metric loop the moment they went from "growing" to "compounding." At 2,399 signups and 56 completed rentals, lend-it is at the exact stage where the loop starts paying rent.
The Loop
Two metrics · Two arcs · One compounding cycle
1 · Demand Signal
Borrower Search Data
Every rental query — the ones borrowers type on lend-it, and the ones they type everywhere else. Two tiers of the same signal.
Primary· On-platform
Every borrower action on lend-it itself. Each unmet query names a specific lender that hasn't been signed. Three streams together:
Search log with results_count and zero-result flag
Inquiry count and timestamps, per listing
Checkout and booking abandonment events
"342 people searched 'bounce house rental Ottawa' and found nothing."
Secondary· Market-wide
Population-scale intent — what Canadians want to rent before they find lend-it. Sizes cities and categories before we spend an outreach hour. This is the layer that surfaced Montreal's French-language demand: +325% over English-only, invisible to any competitor running English-only SEO — and ranked all 15 Canadian metros for the rollout.
Sources Google Ads Keyword Planner · Google Trends · Google Search Console · TikTok search · SerpBear
2 · Revenue Signal
GMV per Lender
Gross rental dollar volume per lender — 30-day and lifetime — sliced by category and city, with platform take separated out. The dollar receipt of every listing decision, per person.
"Order count treats a $20 drill rental like a $500 party bundle. GMV per lender is the only view that answers who to keep, upgrade, or replace."
Drivers· What moves the number
GMV per lender doesn't move by itself. Three inputs, all measurable on-platform, do most of the work:
Sources platform orders · payment processor (Stripe / Moneris) · listing metadata for category + city
The Playbook
Two metrics have carried every marketplace that scaled.
The pattern isn't lend-it-specific — it's what happens in every two-sided market. The demand signal reveals where supply is missing. The revenue signal reveals which supply is actually working. The winners obsessed over both, in that order, on repeat.
Marketplace
Their demand signal
Their revenue signal
Airbnb
Wishlist saves in cities without listings
Revenue per Superhost
Uber
Unfulfilled ride-request heatmaps
Weekly earnings per driver
DoorDash
Cuisine searches with no restaurant match
Monthly GMV per merchant
Etsy
Search terms without matching listings
Top-seller GMV rankings
Turo
Booking searches in empty cities
Trip revenue per host
Lend-It
Rental searches returning zero results
GMV per lender, by category + city
Why the pattern is universal
In every two-sided market, revenue resolves to active buyers × conversion × transaction value × take rate. There are exactly two variables inside that equation you can grow at will: add supply that matches unmet demand, and extract more from the supply you already have.
Everything else — response time, listing quality, retention, referrals, insurance attach — is a knob attached to one of those two levers. Real, worth turning, but only after the lever exists.
Focus these two and the equation moves. Focus anything else and you're turning knobs with no lever behind them.
How lend-it runs it
Each rotation of the loop sharpens the next.
Each pass improves the input to the following pass — which is what makes it compounding, not a checklist. Miss any step and the loop degrades to a one-off campaign.
1
Search reveals a demand hotspot
Zero-result queries name the exact city and category with unmet demand — the highest-signal recruitment brief you can get.
e.g. "342 zero-result queries in Ottawa for event tents last month."
2
Recruit lenders to fill it
Sales targets the exact gap, not a generic pool. Conversion on this outreach is dramatically higher — you're bringing proven demand, not a pitch.
Sign three tent-rental businesses in Ottawa this week.
3
Listings go live, GMV starts flowing
New lenders begin generating rentals in the category and city that was empty. Every dollar earned now carries a lender name and a source query.
Category-level GMV per lender emerges within 30 days.
4
GMV data sharpens the next round
Which lender profiles convert best? Which cities are top-heavy vs long-tail? Which categories punch above their query volume? The answers feed straight back into step 1.
Recruit the profile that produced the top 20% of GMV.
This Week
Two moves to make the loop real.
1 · Instrument
Log every search event with query, city, category, and results_count. Surface the zero-result view as its own report. That report is the recruitment shopping list.
2 · Surface
Expose GMV per lender (30-day + lifetime) in the admin. Sort descending. The top 20% is your best-lender profile; the zeros are your churn queue.