[saas_co_de]

It does seem that there are glaring flaws in the methodology:

1) They use average numbers for vehicle expenses which are going to be much higher than what professional drivers with strong profit incentives are going to pay

2) They report on a sampled average of all Uber drivers while acknowledging that the vast majority of these are part time and temporary employees who are not really optimizing their profits. If they looked only that those who do Uber full time or for more than temporary work they would get a completely different story.

[ejstronge]

I’m not sure what I would do differently - what would you have done to address those two problems? Using averages and talking to a representative sample of all drivers seems like a standard approach.

[saas_co_de]

1) Using one methodology for costs and a different one for revenue is a huge red flag. If you are going to use a driver survey to revenue then you should use the same methodology for costs.

2) The researchers acknowledge that the population distribution is highly skewed (more than 80% of drivers are less than 40 hours per week) with a large group of low-mile drivers on one side and a small group of high-mile drivers on the other side.

When the distribution is extremely skewed like this the median does not provide a very meaningful picture of the population.

The person in the middle who makes $3/hr might still only be driving 10 hours per week.

Professional taxi drivers typically work 12 hours per day, 6 days a week or more. NYC actually just recently passed laws to set 12 hour days and 72 hour weeks as a maximum because people were working over that resulting in safety issues. (http://abc7ny.com/traffic/nyc-putting-limits-on-cab-driver-h...)

If over 80% of Uber drivers are working less than 40 hours per week then less than 20% are working anywhere close to professional taxi driver hours but that group is probably providing the majority of ride miles and getting an even bigger majority of revenue and profit because they optimize their revenue.

For instance, part time drivers will drive in their spare time, while professional drivers will drive at peak hours, which makes a huge difference in revenue.

[ashelmire]

I'm with you, this study isn't worth a damn. I have friends that make $200-300 per night (in maybe 8 hours) driving for Uber (after Uber's cut). They'll use, what, $20-30 worth of gas in that time? And the rest of the costs (insurance and vehicle cost/wear namely) are negligible (tax deduction does not indicate real cost). Driving revenue in cities is way off the right of their chart. Seems like just another contribution to bunk "science".

[pwinnski]

That people think vehicle wear costs are negligible is how Uber manages to convince people to drive for such low wages. People are bad at math.

[ashelmire]

If someone is making $20/hr net income driving at night (pretty average for big cities according to this: https://drive.google.com/file/d/0B1s08BdVqCgrZWZkV0ZfZnhGUGc...), do you really think the wear costs are more than a few bucks per hour?

[wgerard]

Usage costs for cars are usually tagged at around $0.50-$0.60 per mile [1]. Looking around, seems like the average 8 hour uber driver is putting on about 100-200 miles per shift [2].

So assuming 150 miles per 8 hours, the average uber driver is paying about $11/hr in fuel and depreciation costs.

So the net income after that would probably be more like $9/hr.

1: http://newsroom.aaa.com/tag/driving-cost-per-mile/

2: https://www.quora.com/How-many-miles-does-a-full-time-driver...

[ashelmire]

The analysis I linked above calculated the net earnings after all expenses as being ~$13/hr on average (and included all registration and insurance costs as well as the AAA figure in expenses). That would mean higher hourly earnings for part-time drivers in big cities (who often only work Friday/Saturday nights).

[wgerard]

Hm interesting, the report suggests that 8-hour drivers are doing more like 60 miles per shift (($4.79 * 8) / 0.6).

Curious what the huge discrepancy is there (between self-reported and gathered data), though without being able to see the underlying data I guess we can't tell.

[JackFr]

The results also seem to defy common sense.

While it's relatively easy to believe that the income from driving for Uber and Lyft is lower than most people imagine and it is relatively easy to believe that a small fraction of drivers are actually losing money by driving, it is difficult to believe that that fully 1/3 of drivers are continuing to impoverish themselves.

I'm not saying it's not true, but without supporting details it seems clickbaity.

[notfromhere]

Most people aren't math savvy and the long-term costs of riding for Uber are more 'invisible' than the short-term cash that you'll be getting from driving. Uber/Lyft functions by letting car owners get a quick buck in exchange for the long-term value of the car.

People use payday loans even though they're fairly scammy and financially irresponsible, why would Lyft/Uber be any different?

[rileymat2]

I don’t see why this is hard to believe, according to the article it claims the turn over rate is high.