The market is convinced building these huge treasuries of consumer data, at any expense in cost and creepiness, will result in wildly better advertising performance. We've given Google and Facebook 12-digit valuations on that principle.

But advertising is a 80/80 problem: You can get to 80% of potential results easily, and spend 80% of your money, effort, and goodwill chasing the rest. Content-based ads (think the early era of Google AdSense when the ads would be text or fall back to PSAs) sell pretty well, considering you didn't need an exabyte of profile data to build them.

Eventually, fundamentals have to come into play-- these companies need the revenue numbers to justify the valuation-- and the only way to get there is to charge premium prices for their advertising services. At that point, the music stops, when advertisers realize the cost-per-acquisition is poor compared to much cheaper, more scattershot technology, or even just earlier stops on the targeting continuum.

Logistics is a low margin winner-takes-all industry which is fairly well understood.

Just saying Facebook had 10x the user as Uber is meaningless.

I feel like Uber's biggest revenue generator would be putting a small smartboard in front of passengers and selling targeted ads.

Data isn't valuable. Using that data to create a product that people will pay for is (e.g. showing an ad to someone en route to the restaurants across the street).