A Quick Note on Airbnb’s San Francisco Report

A bad-tempered post, for reasons that will become apparent. The interesting bit is the map near the end.

So today Airbnb came out with a report to study the company’s economic impact in San Francisco. Or rather, as per usual, they didn’t: they talk about a report but don’t actually show it, which is par for the course with them. One has to wonder why.

So in their blog post here is what they say, and here are a few other questions they should answer and observations. I do hope people interested will talk to others who track what effect Airbnb has in their city, such as Share Better San Francisco.

Airbnb Claim: The Airbnb community contributed nearly $469 million to the San Francisco economy last year.

Response: What is  “contributed”? Does this mean “what visitors paid in rent”? or “what visitors spent while visiting” (and if so, how do they estimate it). Does “the Hilton Hotels community” contribute the total hotel income it makes from hotel rentals? They really should say what they mean.

Airbnb Claim: The average Airbnb host earns $13,000 per year hosting – money they use to pay the bills and stay in San Francisco, and shop at businesses like yours.

Response: They say the “average” Airbnb host, so I’m guessing this is the mean income, not the median. The mean will be more than most hosts earn, so the number is higher (better) than if they chose the median. (And if this is not what they are doing, why don’t they say?)

Airbnb Claim: The Airbnb community supports 3,600 jobs at the local neighborhood businesses they patronize.

Response: supports? what does “supports” mean?

Airbnb Claim: 72% of Airbnb properties are outside of traditional hotel districts, in neighborhoods that haven’t benefited from tourism in the past.

Response: This is what they always say to show how diverse they are. But looked at another way, my estimates show that over half of the total visits happen in just a few central San Francisco neighbourhoods (which have definitely ‘benefited from tourism in the past’, as shown here. Not quite as diverse as they claim.


Airbnb Claim: The typical Airbnb property is booked about 6.5 nights per month, underscoring the point that these are people who are simply sharing space in the home in which they live.

Response: Note that they say “typical”, presumably meaning “median”, which will be lower than the “average”, so thay can emphasize the occasional nature of their hosts. It’s a slanted picture. Or, again, if it isn’t, why don’t they say so?

They could also point out that over 2/3 of their income comes from people renting out whole homes:



or that over a third of their business comes from people with multiple listings.san_francisco_multilister

But of course they don’t.

What’s going on with Airbnb in LA?

Today’s LA Times says “Airbnb cuts ties with vacation-rental firms in Los Angeles“. The story says that “Two of the home-sharing giant’s largest Los Angeles-area hosts—vacation-rental firms with dozens of apartments apiece—said Friday that Airbnb had dropped them from its site this week, canceling upcoming bookings and scrubbing their listings.”

An Airbnb rep called one of the affected landlords (AE Hospitality) and “mentioned the growth plans of Airbnb conflicts with us listing on their website… No explicit reason was given.”

I got a call from Tim Logan, one of the reporters on the story, asking if I could see anything going on, and was able to take a quick look and (at that time) it looked like 10 of the top 13 hosts (in terms of number of listings) were completely removed from the site, and I’m quoted in the story with that figure. Now I’ve been able to take a more complete look and can say a bit more.

There are two questions, of course. One is “What is happening?” What listings are being removed, whose are they, and what can we tell about those listings? The second question is “Why?” Let’s start with the “What?”

I ran a survey of Los Angeles in October 2014, and another over the last couple of days. Here are the main findings.

There is always a lot of flux in Airbnb listings, and much of that comes from people who give the service a try and then decide it’s not for them. But in Los Angeles there is something else. Here is a table showing the top   hosts (by number of listings) from October 2014, and the number of listings they now have on the site.

Table 1: The top hosts (by number of listings) in October 2014, and the number of listings they had then (2nd column) and have now (3rd) on Airbnb. The link goes to the host page (if it exists) or is redirected to a generic Airbnb page if the host has been removed.
Host ID October 2014 April 2015 Reviews Average Rating
3965199 74 0 15 4.50
1463129 73 0 397 4.47
558295 50 0 113 4.54
293855 30 0 296 4.26
7954 26 0 47 4.60
295792 23 0 143 4.68
3281907 23 0 25 4.50
66892 22 0 65 4.68
2917744 22 1 291 4.85
4925293 20 0 27 4.85
9088345 20 0 120 4.41
465624 18 18 789 4.44
8608492 18 0 39 4.42
3392276 16 18 43 4.67
8518444 16 21 23 4.52
49765 15 13 1 0.00
148358 15 11 160 4.95
1497543 15 15 252 4.76
3127239 15 15 184 4.32
8206200 15 15 109 4.39
57161 14 16 145 4.31
1466173 14 15 517 4.30
3324376 14 15 101 5.00
11907135 14 0 39 4.18
1316725 13 14 9 4.83
1458653 13 13 291 4.60
3947940 13 13 142 4.23
7266554 13 24 127 4.58
16151602 13 14 67 4.34
489690 12 12 46 4.72
1173362 12 0 150 4.45
2324191 12 15 173 4.32
3967730 12 14 118 4.64
4265938 12 8 42 4.11
5592355 12 10 68 4.62
219283 11 16 34 4.78
239712 11 12 590 4.23
1221532 11 12 25 4.30
1254174 11 12 86 3.90
3097566 11 11 88 4.45
9762413 11 12 253 4.57
9928881 11 0 20 4.40
19341612 11 0 12 4.54
213865 10 9 158 4.46
2622454 10 10 780 4.52
5281327 10 4 212 4.65
5425196 10 0 190 4.17
9505402 10 19 247 4.65
9759851 10 11 8 4.69
3343353 9 14 14 4.89

Airbnb has basically removed all the top hosts (who had 20 or more listings) and also removed a smattering of others. (If you have a booking at this listing I would make other plans.)

So why has Airbnb done this? They made their usual bland, substance-free comment to the LA Times: that the company’s “mission is to connect hosts with guests and provide a quality, local and authentic experience. We routinely review our platform for market quality and adherence to this mission.”

Quality looks out: as a user to the site what you see is the rating of the listing, and many of the eliminated hosts have high average ratings. For example, the Urban Flat Team have had their listings removed but had an average rating of 4.85 (out of five), while the Venice Beach Hostel has a rating of only 3.90 (which is pretty crap) but is still on the site.

It’s also nothing to do with the type of listing. Most of the top listers are listing out mainly “Entire home/apt” listings (as opposed to a private room in a house or a shared room) but then so does Prive Luxury Rentals and they have actually increased their number of listings from 16 to 21.

What’s left is the statement of the sales rep to the AE Hostel agency: that these listings are interfering (probably with no fault of their own) with Airbnb’s public image. The LA Times mentions the possibility
of maneuvering in advance of an IPO, and this must be in the minds of some of the Airbnb executives. Also, the LAANE report from last month may be having an impact.

My guess is that Airbnb knows it has to look after its image as a peer-to-peer company, and that they are removing some of the most high profile professional partners from the site in preparation. So all well and good (although hardly fair on those partners), but is it sincere?

Probably not. Here are a few facts that Airbnb will probably not publicise:

  • In October 2014 there were 49 hosts with 10 or more listings in LA. Now there are 60.
  • In October 2014, visits to Airbnb hosts in LA were split evenly between “single-lister” hosts and “multi-lister” hosts. Now the balance has tilted slightly towards multi-listers: their percentage has increased from 50% to 52%.
  • In New York, where Airbnb previously had a clean-up of large listers, the number of hosts with 10 or more listings has increased from 19 to 46 between October 2014 and March of this year.

In short, the professionalization of the Airbnb business continues, and this looks like just an attempt to remove some of the more obvious commercial companies.

Could it be something else? Of course it could, but if Airbnb wants to be taken seriously as a company that can regulate itself (and it really does) then it needs to come clean with the reasons for its actions rather than hiding between vacuous PR speak and turning, seemingly at random, on its partners.

Yesterday in Airbnb

First, this report on Airbnb in Los Angeles, by the Los Angeles Alliance for a New Economy (LAANE), is fantastic, especially compared to Airbnb’s own study from December. It mixes data with thoughtful commentary with investigative work that could only be carried out by someone who knows the city. (I did contribute a bit of help in the data collection area, but had nothing to do with writing or the report itself.) Here are some highlights.

  • Airbnb’s report has this to say about its impact on neighbourhoods (in total):

Airbnb distributes economic impacts to neighborhoods that have not traditionally benefited from tourism spending. With Airbnb properties in more than 80 Los Angeles neighborhoods, Airbnb visitors are staying in and exploring places they might never have otherwise visited.

The LAANE report shows that the Airbnb business is, in fact, concentrated on areas that are already heavily affected by tourism: while “AirBnB has units listed throughout Los Angeles, but just nine of the City’s 95 neighborhoods are responsible for generating 73 percent of the company’s revenue.” In tourist hotspot Venice, one in eight units are now Airbnb rentals.

  • 1010 Wilshire is a high end apartment building with 227 units in Downtown Los Angeles; but 20% of its units are now listed as tourist accommodation on Airbnb.
  • The host you see on the site, with a photo and personal note, may not actually be the host. “Danielle and Lexi” were two young women with “verified ID”, but it turns out that they are just a front for Ghc vacation property rentals.

And there’s a lot more. Worth a read if you have any interest in the subject. The report has been picked up by the LA Times, Curbed LA, and more.

Elsewhere, Inside Airbnb has extended its beautiful maps to Portland. They have also collaborated with Willamette Week to chart the shape of Airbnb’s business in the city, and the results will be familiar to readers of this site: Airbnb’s story of being a site where “regular people occasionally rent the home in which they live” is misleading.

The real Airbnb story is gradually getting out there, of a site that uses the heartwarming stories of a few of its hosts to provide a cover for a growing cadre of professional renters using the site to avoid municipal regulations around safety (no fire inspections), zoning (driving gentrification of tourist areas), and taxes. It’s good to see diligent, talented people like LAANE and Inside Airbnb push the debate forward.

Going quieter for a while

This blog will be going even more quiet than usual for the next few months. The reason is a good one, which is that I have a contract to write a book on and around the sharing economy. As the writing has to be done in and around my actual job and home life I will have to put any and all available time into it for a while.

The Airbnb coding and analysis turns out to be a great way to procrastinate from actually writing, so I will have to put it to one side for now.

In the meantime, anyone interested in the issues raised here may like to look at the following:

  • Henry Farrell’s “Dark Leviathan” in Aeon Magazine is one of the best things I’ve read in ages on issues of trust, governance, and digital technology. Plus it’s got lots of juicy details on Silk Road.
  • Murray Cox has put together a beautiful site called Inside Airbnb, which maps the New York data in more detail than I do here. It’s invaluable. Plus it has led to other analyses, like this one here by Ben Wellington, which delve further into the specifics of New York.

See you back here before too long, I hope.

Roman Holiday: Airbnb and Gentrification

I’m a very lucky person.

For example: while I live in Canada, my birth family lives in the UK, and while I was visiting them last year my sister and I got to take a trip to Rome for a few days. It was a great trip (we stayed in a rented apartment, in case you were wondering, but not rented through Airbnb), and one of our favourite parts of it was spending an afternoon in the Trastevere district. While Trastevere doesn’t have the big sights, it has many cobbled streets, charming restaurants, a beautiful square, and one of the oldest cathedrals in the city. It’s a more artisan, bohemian atmosphere than some of the other areas of the city, and that’s part of its appeal.

So like many other visitors we loved Trastevere, and like other tourists we brought money with us, and spent some of it there, so in a way our visit was good for the city too. But of course our presence is a mixed blessing: too much tourism can erode the very things that make Trastevere so atmospheric and special — especially when those things are the “authentic” life of the locals — raising the cost of living, and putting pressure on property prices. In the end, as tourists you have to hope that the city of Rome and the neighbourhood itself find ways to balance the various pressures acting on the place where they live.

Not surprisingly, Trastevere does have conflicts over gentrification and the impact of tourism. Here‘s a story from last September. In 1956 a cinema opened in Trastevere called Cinema America, and it was a feature of the area until it closed, sometime around 2000. In 2004 it was bought by a new set of owners who literally wanted to pave it and put up a parking lot, along with some apartments. There was local opposition to the plan, and stalemate until in November 2012 a group of young people occupied the cinema. By all accounts, they turned it into a focal point for the community and also attracted some big names of Italian cinema and the arts. In September 2014 the occupiers were evicted by the police, but they have apparently set up elsewhere in the neighbourhood and the campaign continues.

What does all this have to do with Airbnb? Well, this…

Here’s a map of the 8,000 Airbnb rentals in Rome as of last May, coloured by price (red is the most expensive, then yellow, green, blue and violet). You can see that the rentals are most dense and most expensive in the middle of the city.

rome1I use the number of reviews as a proxy for the number of visits, and so I can estimate which listings are the most valuable to Airbnb. Here is (by this estimate) the most valuable listing in Rome — the one with perhaps the highest rental income of any in the city. It’s the red dot in the map below.

rome2If you look just south of this place, you can see “Basilica di Santa Maria in Trastevere”. This listing is, you will not be surprised to know, right in Trastevere. So what kind of a place is it? Well, here is what it’s listing page looks like today (January 2015).


It’s a beautiful looking place, no doubt — although not exactly cheap at $700 Canadian per night. The listing text says that it has been “Selected by the airbnb staff who stayed and filmed here for the upcoming airbnb hosting in Rome video” so its charms have been noticed by the company too.

Airbnb, of course, says that its host “are regular people who occasionally rent out their homes and use the income they earn to pay the bills.” So who is the host for this particularly successful rental?

Well, it turns out that Martin is not so much a regular Trastevere resident. In fact, he is a Harvard-educated tech entrepreneur from Austin Texas who is “rent[ing] out the places I bought with the proceeds from my last software company.” He is now CEO of Vreasy, a “novel software platform” which is “a growing force in the tourism and travel market.”


Martin is listed as an Airbnb “superhost”. In addition to the Historic Nobleman’s Loft he has four other listings on the site – one more in Rome (an “Ultimate Panoramic Roman Penthouse”), two in Barcelona, and one “Seaplane cabin” that he will fly to any legal European lake so that you can use it as a hotel.

Now I don’t know Martin and have made no effort to talk to him, because this is not really about Martin. It’s about Airbnb, and the gap between the sepia-tinted “regular people” they often talk about, and the reality of multiple-property owners. It’s the gap between their message of caring for neighbourhoods and the reality of gentrification driven by uncontrolled tourism, which they are playing a big part in accelerating. It’s the gap between their claim to care, and the reality of callous profiteering. Airbnb’s true business in Trastevere seems to be gentrification personified, and it’s exactly the kind of business that the locals of Trastevere are protesting.



The Sound of Silence, or What Uber and Airbnb’s Missing Data Tells Us

Episode 1: May 2014

Last May, Uber wrote this in a blog post:


UberX driver partners are small business entrepreneurs demonstrating across the country that being a driver is sustainable and profitable.  For example, the median income on uberX is more than $90,000/year/driver in New York and more than $74,000/year/driver in San Francisco.

There are all kinds of issues with the claim, but put them aside for the moment. Instead, here’s a question: assuming the claim to be true, what does Uber’s post tell us about the income of uberX drivers in cities other than New York and San Francisco?

One obvious thought is that it tells us nothing, because it only makes claims about NYC and SF. Another is that we could extrapolate from these numbers as if they were randomly chosen, and guess that the income in other cities is somewhere in a distribution centred around $82K. That’s what the Washington Post did by implicitly comparing the Uber numbers to a national average for taxi drivers:

Estimates of the typical cab driver’s salary hover around $30,000. According to Uber, the median wage for an UberX driver working at least 40 hours a week in New York City is $90,766 a year. In San Francisco, the median wage for an UberX driver working at least 40 hours a week is $74,191.

But Uber’s blog post actually tells us something different about driver income in other cities: it tells us that it is less than $74,000. How come? Simply because this is a blog post from Uber. Uber wants to make a compelling case to the media and to the public, so if drivers in Chicago or Los Angeles were making more money than those in NYC or SF they would have appeared in the post instead of those two cities. So we don’t know how much drivers elsewhere make, but we can be pretty sure that it’s less than $74,000.

Episode 2: Oct 2014

In October 2014, Uber published another blog post saying that drivers in New York City take an average of $36.16 per hour in fares, and make $25.17 per hour after Uber’s cut, but before expenses. (That’s about $50,000 per year for a 40-hour week, by the way.) What does this tell us about driver income in other cities? Again, it tells us that it’s probably less than $25 per hour because otherwise they would not choose New York City for their blog post.

Episode 3: January 2015

So last week Uber posted both a survey of its drivers and a report (PDF) by eminent economist  Alan Krueger and Uber Head of Policy Research Jonathan Hall. The main things to come out of the report were that Uber is growing really quickly – impressive, but no surprise — and that Uber drivers are paid better than taxi drivers. That conclusion comes from Table 6:  uberwages

The table confirms what I was saying above: earnings in New York City are 50% more than anywhere else apart from San Francisco, and San Francisco is the second highest earning city.

As several other commentators have pointed out (eg Ellen Huet at Forbes, Jacob Davidson at Time, Andrea Peterson at the Washington Post), the comparison between Uber “earnings per hour” and taxi “hourly wages” is apples and oranges, because the report has nothing to say about Uber driver expenses. Krueger and Hall have this to say in their defence:

A detailed quantification of driver-partner costs and net after-tax earnings is a topic of future research. Nonetheless, the figures suggest that unless their after-tax costs average more than $6 per hour, the net hourly earnings of Uber’s driver-partners exceed the hourly wage of employed taxi drivers and chauffeurs, on average.

(Oh Alan Krueger: you were given a set of data and “full discretion over the content of the report” but you come out with this? I’m not angry, I’m just disappointed.)

So what do we know about driver expenses? The Sound of Silence rule tells us that, if expenses made a good story for Uber, the company would have included expense data in the data set given to Krueger. In fact, if expenses made a good story, they would have been included in the earlier company posts as well. So we can be pretty sure that the expenses of drivers are $6/hour or more in most cities.

Is $6/hour expenses reasonable? Certainly: the Washington Post’s Peterson notes that the IRS estimate of costs is 57.5 cents per mile, but that Uber don’t include distance driven in their report. Still, 12 miles an hour (including driving to pick up a fare and driving back to a good spot afterwards) is not implausible. And then there is the thorny question of commercial insurance, which would be an additional cost. It seems likely to me that Uber drivers are getting about the same as taxi drivers.

Episode 4: January 2015

Just to finish off, here is another sound of silence. Thanks to the efforts of Share Better, my data on Airbnb rentals got picked up in New York City last week when City Council held hearings about short term rentals. That data suggests that the percentage of illegal rentals in Airbnb’s New York listings has decreased only slightly since a year ago, and that the decrease has been swamped by an increase in the overall number of listings. There is not enough firm data on the Airbnb public web site to be absolutely solid on this, but that’s how things look.

What did Airbnb have to say? Their response is that the data is “flawed” and “inaccurate”, and that is literally the extent of what they have to say. So it’s kind of a joke. This piece in Mashable by Jessica Plautz includes Airbnb saying they don’t really know anything about their hosts, and Josh Dzieza in The Verge includes more of Airbnb’s exchange with the council in which the company says, well, nothing.  Airbnb executive David Hantman’s testimony ‘acknowledged that his company did not keep track of the number of users who are renting out their apartments illegally. “We don’t research that,” he said during the hearing.’

Sometimes I wonder if I’m doing the right thing going after Airbnb, because there is a role for occasional casual short-term rentals. Maybe the company’s heart-on-its-sleeve attitude is for real? But then I realize that in the last year they could have put out realistic, convincing numbers at any time (and saved me a lot of work) but they haven’t. And like Uber’s silence on driver expenses, Airbnb’s silence speaks volumes: if the data were good for the company, they would have made them public. The fact that they haven’t (and their content-free responses to people trying to piece together a picture of reality) makes me more convinced than ever that they have things to hide.