I Thought My Uber Driver Got Half

daily drive

Jun 19

Is Uber’s algorithm its greatest strength and its greatest long-term risk?

My Passenger

He climbed into my car wearing a walking boot and laughing at himself. He had gotten off work, stopped for a couple of beers, and I was taking him home. The injury, he explained, came from trying too hard during a family game of wiffle ball. His cousins told him to try, so he hit five home runs in a row. Then he decided to hit one into play. Now he actually had to run, and somewhere between home plate and first base, he ended up with a stress fracture.

It was a funny ride, the kind of ride I enjoy. Two strangers talking about life. It’s the kind of ride that reminds me why I enjoy being an Uber driver.

But then he started talking about money.

He told me he had paid $20 for the short one-mile ride. Then he said, “And you’re probably only getting $10.” Like most passengers, he assumed I was getting at least half. Ten dollars seemed reasonable; the bare minimum. That’s what most passengers tend to believe.

When I told him I had received $8, he stopped.

“That’s F_____ up!”

He wasn’t angry. He wasn’t an Uber driver. He wasn’t trying to make some grand statement about capitalism or labor. He was a guy with a broken foot trying to get home. Yet instinctively, he thought that if he paid twenty bucks to go a mile, the driver ought to get half. That seemed fair.

Then, what felt more like pity than gratitude, he tipped me $6 to make things right. He didn’t tip because I had bottled water or because I drove a clean car. He didn’t tip because I performed some extraordinary act of customer service. He tipped because he thought he ought to. In his mind, the tip wasn’t gratitude. It was a correction.

The Problem

I’ve driven more than 15,000 Uber rides over the last six years, and I’m happy being an Uber driver. I like the instant money, the flexibility, being out in the city, and meeting people. The work has supported my family and allowed me to pursue other things I care about, like writing this essay. Uber driving is the best job I’ve ever had!

But being grateful doesn’t mean everything is perfect.

I’m not writing this because Uber is evil. In fact, I’m a bald-eagle, stars-and-stripes, flag-waving American free-market capitalist. I love investors who take risks and provide capital to build profitable businesses. I support people getting rich because who knows, maybe I’ll get rich one day. I don’t believe Dara Khosrowshahi wakes up every morning trying to make drivers miserable. I want Uber to succeed. What they’ve built is remarkable.

But I also think there is a problem. And the more I study it, the less greed is the problem, and the more complex it is.

Transparency and Complexity

Years ago, Uber operated under a rate card. Every driver in a city earned a fixed amount per mile and minute. It wasn’t perfect, but everyone understood the arrangement. (Note: I don’t support returning to a rate card, because not all drivers and not all cars are the same.)

But upfront pricing, introduced several years ago, changed that.

Today, the passenger’s fare and the driver’s pay are no longer connected. The passenger sees one number. The driver sees another. Uber sees both. Somewhere in the middle sits an algorithm determining how much the passenger is willing to pay and how little the driver is willing to accept.

That isn’t sinister. Airlines do something similar. Hotels do something similar. Heck, universities use dynamic pricing to extract as much as possible from wealthy alumni’s kids to fund the low-income biology scholarship. Dynamic pricing itself isn’t the issue.

The issue is that nobody seems to understand the math anymore.

Last year, I conducted a study on 159 identical rides between Ithaca and Syracuse Airport. Of the 159 trips, I narrowed it down to 129 with complete data. Same driver (me). Same car (my Tesla). Same route. What I found fascinated me. Passenger fares varied wildly. The same sixty-two-mile trip ranged from roughly $79 to $168. Insurance fees fluctuated dramatically. Take rates moved all over the place. And the more data I collected, the harder it became to explain.

One data point that was hard to explain was commercial auto insurance. Sometimes Uber said the insurance was $13.75, and other times it was as high as $50. But it was the exact same 62-mile trip, with the same driver using the same car. This is the same thread that we’re all pulling on.

My friend Professor Len Sherman of Columbia Business School has been working on this same problem in parallel with me for a while now. His recent article is the data-rich version of what researchers have been seeing. He analyzed nearly 50,000 trips from three veteran drivers and concluded that Uber’s real take rate, the kind people used to mean when they said “take rate,” has climbed from about 15% to more than 50% over the past decade. He even ran the numbers on my own Ithaca–Syracuse trips. Consumer Watchdog, Consumer Reports, Business Insider, and other recent investigations suggest this confusion isn’t isolated to me.

Passengers don’t understand pricing. Drivers don’t understand compensation. Consumer advocates are trying to understand it. Investors argue about it. Researchers like Len are doing everything they can to solve it.

Maybe that’s the story. Not greed. Not exploitation. Trust.

Trust and Brand

One thing keeps coming back to me. My passenger didn’t think he was overpaying. Yes, it was expensive, but at least I was getting half, he assumed. Only when he learned otherwise did he feel obligated to fix it. His tip wasn’t a reward or gratuity. It was an attempt to restore some sense of fairness.

And I wonder how often that happens.

Passengers paying $120 for an airport ride also assume the driver must be making $80. But in reality, the driver might only receive $40. Or worse, after the driver’s expenses, as little as $25. If the driver spent thirty minutes waiting for that ride, another thirty minutes driving back empty, and seventy thousand miles a year maintaining the vehicle, the picture becomes very different; Uber driving becomes a minimum-wage job.

Here’s the part most people miss. If a passenger Googles “how much does Uber take,” they’ll find 20%. Uber’s PR response to the Consumer Reports research states:

Uber’s U.S. take rate is around 20%: The report’s findings on take rate were distorted by a flawed experimental design. Volunteer riders and drivers were placed in the same room—but driver earnings depend in part on how far the driver must travel to reach the rider. By minimizing pickup distance, the experiment created an artificial scenario that isn’t representative of reality. Adjusting for insurance costs, which we voluntarily carry on behalf of drivers, our average revenue per trip is around 20% in the U.S. and has remained at that level for years.

Okay. But here’s what fascinates me. The more I’ve studied my own numbers, the more I’ve realized there isn’t really one take rate. There are several.

You can look at an individual fare.
You can look at a week.
You can look at a year.

And depending on which math you use, you arrive at very different answers.

Take my passenger’s ride. He paid twenty dollars. I received eight. From his perspective, I got forty percent. Uber would say their actual take was much lower because commercial auto insurance and other costs have to be included. And they’re right.

But then look at a week. Last week, I received forty-three percent of the passenger fares. Uber’s service fees accounted for roughly 22%, which is remarkably close to the 20% figure the company publicly reports. But that leaves another twenty-three percent for commercial auto insurance and customer discounts.

Then look at the year. Last year I completed more than 3,200 trips. My annual tax documents combine service fees and commercial auto insurance into a single category. Suddenly, the distinction Uber draws between take rate and insurance disappears because they’re combined on the documents I use to prepare my taxes.

And that’s what fascinates me. Passengers are doing mental math. Drivers are doing weekly math. Researchers are doing trip math. Uber is doing adjusted accounting math. Investors are doing financial statement math. Everyone is using different math.

No wonder people are confused. No wonder people disagree. Because perhaps there isn’t one answer. Maybe there are several. And complexity itself has become the problem.

Which brings me to what I think is the biggest misunderstanding in this entire debate. Uber likes to talk about active hours, and I understand why. Active hours look fantastic. That’s when the work is getting done. That’s when passengers are in the car, and the hourly numbers can appear impressive.

But active time isn’t the whole job.

The whole job includes waiting. Deadhead miles. Positioning. Declining bad offers. Sitting in parking lots. Running multiple gig apps, hoping to fill dead time. Making seven-second decisions while navigating traffic.

Online time is labor too. And that’s the metric professional drivers live by.

Online time is when I’m working.

Uber’s Brand Opportunity

The thing is this: professional drivers matter. We’re the drivers with thousands of trips, high ratings, and years of experience. We aren’t random suppliers floating around the edge of the marketplace. We’re part of the infrastructure. Like it or not, we’re Uber’s customer service, the brand ambassadors. And I wonder if Uber leaders and investors fully appreciate that.

Because somewhere in all of this, someone still has to drive the car. Someone has to maintain the vehicle. Someone has to absorb depreciation. Someone has to transport another human being safely at two in the morning. Someone has to decide whether they’re too tired to keep driving after ten hours. Someone has to handle the in-car customer service.

And we have to do that with a smile on our face.

Those things matter. Tomorrow’s autonomous cars… well… we’ll see. But today still matters.

I don’t think Uber drivers need to become rich. I don’t think Uber owes anyone a guaranteed fortune. But I do think a society that depends on people transporting human beings should care whether those people are making a sustainable living.

Because safety matters. Experience matters. And trust matters.

Drivers and our cars are Uber’s brand.

Complexity Hurts Uber’s Brand

I don’t think Uber has a greed problem. I think Uber has a complexity problem. And drivers, confused and frustrated by Uber’s complexity, hurt the brand.

Many things can be true at the same time. For example, complexity is a genuine advantage for Uber. Yet, I’m also convinced it’s the greatest risk.

But it was never Uber’s greatest strength. The brand was. The service is remarkable. The app is remarkable. For years, Uber was the high-quality, affordable alternative to taxis. That, not the algorithm, is the real asset. And complexity is quietly eroding it.

Complexity allows Uber to optimize. It creates profits. It gives the company tremendous flexibility. But complexity also creates suspicion, and people can tolerate outcomes they don’t like much more easily than outcomes they don’t understand.

Trust doesn’t disappear all at once. It erodes slowly.

Passengers begin wondering why rides are so expensive. Drivers begin wondering where the money went. Tips become wage corrections instead of acts of gratitude. Eventually, both sides begin to feel manipulated, even if no one can fully explain how.

And that’s dangerous, because eventually everyone will have algorithms. Everyone will have AI and mathematics. But not everyone will have trust.

I’m not writing this because I hate Uber. Quite the opposite. I’m writing this because I want Uber to survive. Drivers like me could and should be Uber’s biggest cheerleaders and brand ambassadors!

Because a marketplace where neither side understands the math is fragile. Human beings can endure almost anything if they believe the arrangement is fair. The moment they stop believing that, the whole thing becomes much harder to hold together.

And somewhere, sitting in the back seat with a broken foot after a game of wiffle ball, my passenger understood that instinctively.

He paid twenty dollars; he thought I got ten, but I got eight.

And maybe the most interesting question isn’t where the other twelve dollars went. Maybe the more important question is whether a marketplace built on algorithms can survive if people stop trusting the math. I think that’s the question Uber should care about most.