What YouTube Shopping Taught Me About Mortgage Marketplaces
Arjun Lalwani ran YouTube’s Shopping Monetization team, then got rejected for a mortgage. On why mortgage “marketplaces” are really lead farms — and what an actual marketplace for home loans looks like.
I spent five years at Google building marketplace products — hotel search, then the YouTube Shopping Monetization team, where we launched the YouTube Shopping Affiliate Program and grew a new shopping business from zero. Then I tried to buy a home, got denied by my lender, and had to start the entire process over with another one. The most sophisticated marketplace-building playbook in tech, and none of it had touched the largest purchase of my life.
Why do mortgage “marketplaces” feel like lead farms?
A real marketplace does three things: it aggregates genuine supply, it gives the buyer comparable prices, and it makes money in a way that rewards matching people well. Most online mortgage platforms do none of these. You fill out a form, and your phone number is sold to a set of loan officers who race to call you first. The platform is paid for the lead, not the outcome. That is not a marketplace — it is an auction where you are the item.
What does shopping for a mortgage actually cost the industry?
The deeper problem is that the supply side is drowning in its own process. Originating a single loan now costs lenders nearly twelve thousand dollars — most of it human coordination: collecting documents, re-keying data, chasing signatures, shepherding a file between systems that do not talk to each other.
When I described my own process, the honest summary was that it felt like the 90s: emailing PDFs, chasing paperwork, and bouncing between loan officers. Every marketplace I worked on at Google had eliminated exactly this class of friction a decade earlier.
The marketplace we are building instead
You fill one form out, and we can get a few quotes across a bunch of lenders — the comparison a marketplace owes you, without your phone ringing.
At Ralo we rebuilt the brokerage itself as software. AI agents collect documents once, shop the file across lenders, and coordinate the loan to closing; my co-founder Helly and I run the entire licensed brokerage as a two-person team. Automating the back office is what lets a mortgage marketplace finally work like one: real supply, comparable prices, and an incentive to match you well — because our margin comes from a thin commission, not from selling your contact information.
Mortgages are a marketplace problem wearing a paperwork costume. We solved harder versions of it for hotels and creators. Home loans are next.