An early Uber investor who sold his last startup for $150 million has a new app that takes aim at on-demand services

The biggest problem with “on-demand” apps, the Uber for XYZ model, is that they make whichever service they are selling a commodity, according to tech entrepreneur and investor Scott Belsky.

Belsky, who was an early investor in Uber, thinks the on-demand model is perfect for two crucial areas: driving and delivery.

But for "relationship-based" services — tutors, acupuncturists, chefs, fitness instructors, dog walkers, and so on — Belsky told Business Insider there’s value in finding someone you trust, and can stick with. And he thinks these types of service professionals have been hurt more than they've been helped by technology.

“You don’t need 10,000 hairstylists” on an app, Belsky said, matched to you based on optimal proximity, or user ratings from tons of people you don’t know. You just need a friend you trust, perhaps one with similar taste, to recommend you a good one.

“I like to start with a problem,” Belsky said. In this case, his was that the service professionals he'd had conversations with were dissatisfied with both the old model of doing business and the one provided by new technology.

Here’s how Belsky described it: On-demand apps turn their work into a commodity, online marketplaces (where they can list their services) de-personalize them by focusing on a “star” rating system, and companies that book everything for them take a huge cut of their revenue.

Belsky talked particularly about a massage therapist who he started seeing a few years ago, after traveling back and forth all the time from New York to California took its toll. After a year of seeing her, she told Belsky that the person she used to book clients took 30%.

"I almost choked,” Belsky recalled.

So Belsky started pitching everyone he knew on the idea of a new way to book service professionals, trying to figure out what stuff might work. He got encouragement to keep at the idea from people like Uber cofounder Garrett Camp, and also eventually found the team that would Prefer, including CEO Julio Vasconcellos, the founder of Peixe Urbano and a former member of the Facebook growth team.

But for Belsky, one of the crucial elements in building Prefer was getting feedback not just from techies, but from service professionals of all kinds as well. And it led to some surprises about what they actually wanted out of the platform.

“We assumed they would be excited about superior booking and payments,” Belsky said. “As a client, people are annoyed by the back and forth [of booking appointments]. But it turns out service professionals didn't mind it as much. So much of their time is under-utilized. They don't see it as much as a pain point. The thing that got them really excited was getting referrals to their clients' friends. That’s the Holy Grail for them.”

That aspect became the central piece of Prefer. The app uses your phone’s address book — where you presumably have the numbers of both your friends and the service professionals you use — to build out a network of who is employing whom. You can hide certain ones if you wish, but the point is to be able to see which people your friends have trusted to do certain kinds of work. You can also add testimonials.

“I love this idea of bringing stuff to a small-town mentality,” Belsky said. In a sense, Prefer wants to build you a virtual small town of service providers.

Prefer makes money by taking a fee of every transaction, from 3% to 5%. New clients discovered on prefer platform are in the 5% category, and existing clients are in the 3% one. In addition, Prefer charges a convenience fee for booking.

Belsky’s hope is that this business model will prove attractive for service providers, and perhaps even stretch beyond, into the economy at large.