Can Everything be 'Uberfied?'

We're now entering what I like to call "phase II" of the on-demand economy.  Phase I was defined by the movement's poster child, Uber; and the replication of its model into everything from dog-walking to weed delivery.  The unit economics have made some startups successful, while others flamed out.

But the movement really sprouted from deeper roots. Mobility has conditioned us to treat the smartphone as a remote control for the physical world. Urbanization has created network effect for supply and demand matching. And millennials are a generation for which "on-demand" is practically a tagline.

Meanwhile disaggregated supply, like drivers, has been brought together and made "liquid" by companies like Uber. Given that demand is so pervasive (millennials, etc), the magic is in the supply-side. It's lots of logistics, matching algorithms, and even a legal battle or two.

But the question is where on-demand is headed next? One direction is up the ladder to higher-end professions. It won't just be drivers and dry cleaning, but also lawyers and architects, as we're already seeing. And millennials will be on-demand providers as much as they are consumers.

But where are the limits of on-demand? The question came up recently when watching tech giants like Amazon and Google move into the biggest on-demand category of all: Home Services. We're talking renovating your kitchen, replacing your roof, and unclogging your toilet.

But on-demand home services have proven difficult... just ask Homejoy. The common cry was that its downfall was a weak stomach for impending policy battles around employee/contractor status. But it was really because of the complexities and pitfalls inherent in the vertical.

Though the on-demand model has attractive unit economics, they mostly apply to services that don't have a wide quality variance. Uber works great given a fleet of drivers that are relatively interchangeable. That's makes its supply/demand matching algorithms work so well.

But home service contractors aren't as interchangeable. There are myriad categories and subcategories of service -- which each carry a level of specialization (try looking up window replacement vs. window repair). It also takes place in your home, so trust is huge.

All of these factors make quality control more complex. And complexity in on-demand service equals margin compression. This raises another key challenge of home services: their margins are typically thin to begin with, so there isn't much room for error.

Yet another challenge is "leakage." That's when a customer goes directly to a service provider for the second or third job. It's inherent in home services where relationships are built. But it sidesteps the Googles and HomeJoys of the world, which doesn't bode well in a model reliant on repeat business.

Airbnb fights leakage by providing free insurance and resolving disputes. Google could similarly offer perks to home service providers to keep them on the platform, such as software that helps them manage jobs. Think: scheduling, invoicing, payment processing, etc.

This would essentially be CRM for on-demand workers, which is one of my biggest predictions for "phase II."  That would probably be more fitting to a Salesforce or specialized CRM player, but Google or Amazon could offer it as a retention driver within their on-demand networks.

Meanwhile, Google's scale could be its saving grace in having volume to fulfill all that job specialization. But it might run into trouble in all of the other areas above. If it strikes out, there's still plenty of room in on-demand weed delivery. Maybe there's even an unused letter in its Alphabet for that.