Notes from Personal Democracy Forum 2014

So I was lucky enough to be at the Personal Democracy Forum at the beginning of June in New York. A few scattered thoughts.

I had a prejudice that this would be something of an inward looking conference, and was happy to be proved wrong. The range of perspectives was much broader than I expected, and while the core of the conference is the intersection of the internet with electoral politics and governance, the range of topics was broader than I expected too.

Which made the many references throughout to some loosely-defined “we” a bit odd. These were references from the stage and from organizers that implied a community around the internet and its use in (loosely progressive?) politics. This was the 11th PDF, a lot has changed over the last decade, and I’m not sure there is such a “we” any more.

The highlight of Day 1 was the appearance by video of Edward Snowden, a year after the first published leaks. Snowden impressed. He came across as thoughtful, well-read, articulate, and modest, which is a compelling combination. He even looked bashful at the prolonged applause, which was winning. I wish I could say the same for John Perry Barlow, who was “in conversation” with him, but I can’t. Barlow kept dragging the conversation around to himself and kept prompting Snowden to agree with him – prompts that Snowden politely stepped around. Snowden’s main point was a simple one: that, whatever one thinks of states spying on each other (as in, the US spying on Angela Merkel for example), mass surveillance of entire populations is significantly different and crosses a pretty clear line into illegitimacy. In addition, it has not even proved to be useful for its stated goals of preventing terrorist attacks. Snowden may turn out to be as pivotal figure for our time as Daniel Ellsberg was for the Vietnam War era. The video is here.

Some other selected personal highlights from the talks, with links where I can…

The theme for Day 1 was “Save the Internet”, focusing on the role of an open internet for political freedom. I found Katy Pearce‘s talk on “How Authoritarian Regimes Take Advantage of Social Media” and Zeynep Tufekci’s talk on “Movements in a Connected Age: Better at Changing Minds, Worse at Changing Power?” both fascinating. Pearce has been looking at activism in the former Soviet states for years, and she paints a complex picture of the tensions and continuing changes as digital technologies morph and take on new roles in society. Tufekci returned from her native Turkey with interesting ideas about the difference between what is needed to spark sudden uprisings and what is needed to sustain a social movement or to achieve lasting political change. Demonstrations, she argued, act as a signal for a movement’s organizational capacity. When demonstrations happen more easily, governments may misread the organizational strength they imply–and perhaps this happened in Egypt in 2011–but such misreadings won’t persist.

Later on, after my own panel session on the sharing economy was over (written up by Sam Roudman here, I felt it went pretty well but then I’m not the one to ask), more talks. Susan Crawford on the potential for municipal investment in “dark fibre” (ie, optical cables that are open for use) was all new to me – she writes about it here. Sue Gardner, now leaving Wikipedia, is an inspirational figure and her concerns about Building the Public Internet were compelling. You can find similar thoughts by her here from last year.

Day 2’s theme was “The Internet Saves”, the idea being to focus on digital activities that used the Internet for good. The first morning session was great: Matthew Burton and Mike Bracken both talked about the importance of government service – a welcome change from the usual tech focus on entrepreneurship as the solution to social problems – and both delivered far more than they promised, which is the right way to go. If you have 17 minutes, do listen to Bracken: entertaining and novel. Bracken draws his language from Tim O’Reilly, speaking of “Government as a Platform”, but his conception differs from that of O’Reilly (does one write “O’Reilly’s”? that doesn’t look right). The bulk of the O’Reilly “Government as a Platform” talk is more accurately described as “Government as Data Source”: government provides the data, private industry uses it to build services. Bracken was talking about building a complete stack of services within the government, and to me that’s a different take on the topic, and one that is far more promising.

But then came Brad Smith, Microsoft’s General Counsel, and the contrast could hardly be more stark. Lots of overblown rhetoric and ringing phrases but much less substance and little self-awareness. Let’s just say he sounded like a trial lawyer. And then much of the rest of the morning reflected the Omidyar Network’s approach to dealing with social issues by setting up businesses to deal with it, an approach I am not fond of. Omidyar Network is clearly doubling down: one short talk here was about HandUp, a startup that funnels donations to the homeless. Yes, a startup. So there was more for me to grumble about here, but again – that’s the point of a good conference, to hear different points of view.

The format for the whole program was talks of about 20 minutes with no questions (plus breakouts). It worked well: allowing Q&A would have slowed everything down, and a lot of positions got spelled out just by having the variety of presenters on stage.

So good job TechPresident, and thanks for the opportunity to take part.

The shape of Airbnb’s business (II)

Collecting data for several cities both in November 2013 and in May 2014 gives a look at some ways in which Airbnb’s business is changing in these cities. The cities surveyed are:

  • Paris and New York (Airbnb’s two biggest markets)
  • San Francisco (the home of Airbnb)
  • London, Rome, Berlin (three European capitals and major tourist cities)
  • Toronto and Chicago (two smaller markets).

As in Part I, the surveys were carried out by a two-stage method:

  • Stage 1 used the Airbnb search pages for a city to collect Room ID values (a number that uniquely identifies an individual listing).
  • Stage 2 visited the listing page for each Room ID and collected information about it.

The data were stored in a database and the charts here are based on queries of that database. The survey code (and database) are available on github.

The main observations are:

  • It is possible that the number of listings has reached a maximum in some of Airbnb’s largest cities, particularly in the USA, while other cities do continue to show significant growth.
  • There is a surprising amount of churn in Airbnb’s business. Over a third of the listings on the site in November were no longer there in May, to be replaced by a similar (or, in some cases, greater) number of new listings.
  • Around 90% of all ratings on the site are 4.5 stars or 5 stars (out of five).

These observations suggest some conclusions:

  • Airbnb’s move to professionalize its host base may be a reaction to the churn in the host population, a necessary step to generate continued growth.
  • In April Airbnb expelled 2,000 listings from its New York offerings, claiming that they did not offer a positive experience. Many of those listings must have had either no ratings at all or were rated highly (4.5 or 5) by their guests.
  • The rating system does not provide a reliable assessment of host quality.

Churn

Figure 1 shows the net change in the number of listings in each of the eight cities for which data was collected in both November and May, as a percentage of its November value.

Some cities seem to have maxed out in terms of listings. Six of the cities saw either decreases or increases of less than 10% in the total number of listings. Each of the three US cities in the study lost listings between November and May, while Berlin and Rome were big gainers.

nysuffer.png

Figure 1: Change in number of listings in cities.

The net gains or losses mask a bigger change. Figure 2 shows that in the six month period, well over a third of existing listings vanished from the site, and were replaced by a similar (or, in the case of Berlin and Rome, substantially greater) number of new listings. For both Paris and New York, well over 40% of listings vanished, only to be replaced by new ones. Figure 3 shows the total number of vanished and new listings.

The high churn rate must introduce uncertainty for Airbnb’s future, as it is relying on a steady stream of new listings to replace those that are vanishing from the site. Perhaps this is one of the motivators of Airbnb’s efforts to professionalize its host base: hosts who make a commitment to the business being more likely to stick around.

churn.png

Figure 2: New and vanished listings between November 2013 and May 2014, as a percentage of the November number.

example1.png

Figure 3: absolute number of new and vanished listings

[Note: I did wonder whether this high churn rate could be an artefact: whether the displayed room ID values could have changed between the November and May data collections. A query showed that, of the 10597 listings that occur in both the November and May collections from New York, 10529 have the same host. I concluded that room ID values are stable.]

We can look a little closer at those vanished listings. Figure 4 shows that the vanishing listings have fewer reviews than the overall population in the city.

churn_reviewcount.png

Figure 4: Mean number of reviews for vanished listings and for all listings

Figure 5 shows the ratings that the vanished listings had received from guests. The overall distribution of ratings is investigated later in this report, but Figure 5 shows that, while many of the listings had no reviews at all (Airbnb recently claimed that reviews are left for 70% of visits, so it is likely that these listings had no visits either), of those that did have visits the average rating was overwhelmingly high (4.5–5).

churn_lost_ratings.png

Figure 5: The number of vanished listings by average overall rating

Figure 6 emphasizes the point, showing that the average rating of listings leaving the site is within 0.1 of those that stay on the site.

churn_rating.png

Figure 6: Average rating of listings that leave the site (red) and which have been on the site for the whole period (green). Note that the y axis starts at 4.5.

It seems likely that many hosts try Airbnb, have a few guests (or none at all) and then simply decide that, for whatever reason, the service is not for them. To repeat the conclusion from the overall churn rate, the high turnover rate does raise questions about what long-term population of hosts the site can support, and how it can ensure that those who try and leave the site are replaced by new hosts.

Trust

Trust has long been a key part of the Airbnb business, and technology has been seen as the magic ingredient. Airbnb CEO Brian Chesky expresses the company’s confidence when he says of city-level rules that “they’re primarily set up for screening. To protect consumers. Well it turns out that cities can’t screen as well as technologies can screen. Companies have these magical things called reputation systems… We think government should exist as the place of last recourse.”

There are two sides to Airbnb’s trust system. There is peer-to-peer trust, often described as a reputation system, which consists of ratings, personal profiles, and comments written by hosts or guests and seen by other hosts or guests. The other side is centralized trust, which consists of central payment, verified ID, and a complaint system.

Much of the talk around the sharing economy focuses on reputation and peer-to-peer trust (as in Chesky’s comment above). That’s because it’s the peer-to-peer aspect that is new: after all, in one sense we have forever trusted strangers to deliver services from food to haircuts to regular taxis, and the centralized component of the system is not that different to many hospitality companies. Jason Tanz writes in Wired that

We are entrusting complete strangers with our most valuable possessions, our personal experiences—and our very lives. In the process, we are entering a new era of Internet-enabled intimacy.

So what do the numbers have to say about peer-to-peer trust?

Figure 7 shows the distribution of ratings for each listing. Airbnb provides ratings under a number of categories (cleanliness, accuracy of description etc) and collects them together as “overall satisfaction”. The site does not give the ratings for individuals (although individual comments are visible) but provides an aggregate, from 0 to 5 stars with a half-star granularity.

city_rating_distribution.png

Figure 7: Overall satisfaction ratings for listings in each city. The ratings for 0-3 are so few that they are collected together. Listings with no ratings are excluded.

The vast majority of ratings are either 4.5 or 5 stars. The finding is consistent with ratings on other sites (e.g. an earlier look at BlaBlaCar, or more serious studies of eBay listed there). When we rate each other, ratings become more a courtesy than a judgment. Just as restaurant tips only weakly correspond to the quality of service, so a rating of 4.5 or 5 is more a way of politely concluding an exchange than it is of assessing the behaviour of a host or guest.

The ratings obviously fail to distinguish among the people on the site, and so are not providing the service for which they were intended. This is part of the reason why Airbnb and others have moved to emphasize the centralized trust aspect of their systems. But centralized trust is the same reason we trust the fast-food restaurant cook or the hotel cleaning staff. The major sharing economy sites rely on discipline and the potential for removal from the site to provide security for the users, just as traditional industries do.

Churn and Trust in New York

The Airbnb court case in New York raised the profile of the service there, and resulted in Airbnb ejecting 2,000 listings from its service.

Airbnb claims that “we found that some property managers weren’t providing a quality, local experience to guests. These hosts weren’t making their neighborhood stronger and they weren’t delivering the kind of hospitality our guests expect and deserve. In some cases, they were making communities worse, not better.”

With 10,000 listings vanishing from Airbnb in the New York area, it is difficult to know which 2,000 were removed by Airbnb and which left for other reasons. Some individual hosts with large numbers of listings have been removed, but that still leaves many unaccounted for. Figure ny-vanished shows that, despite Airbnb’s suggestion that the hosts are not “delivering the kind of hospitality our guests expect”, most of the removed listings must have had either no ratings at all, or must have had good ratings from their guests.

ny-vanished.png

Figure 8: The “overall satisfaction” ratings for listings in New York that vanished from the site between November 2013 and May 2014.

The biggest single category of the 10,000 New York listings that have vanished from the Airbnb site are listings with no ratings at all (shown as NULL). After that come the 5.0 and 4.5 ratings, leaving fewer than 800 with bad ratings (4 or lower). For at least 1200 of the listings that Airbnb removed, bad ratings is not the reason.

Combined with the very high number of positive ratings shown in Figure 7, this result shows that the significance of peer-to-peer ratings is exaggerated for Airbnb. While the company proclaims that peer-to-peer rating systems set it apart from old-style methods, it is clear that their trust system is essentially a traditional complaint-based, centralized system. The peer-to-peer ratings may give warm feelings on the site, but Airbnb itself clearly does not trust it when it goes to remove listings from its site.

Conclusions

The data set suggests that there is a surprising amount of churn in Airbnb’s hosts. Over a third of the host population in November left the site, to be replaced by new listings.

It is likely that a large number of people experiment with being Airbnb hosts, and then decide that it is not for them.

It will be interesting to see whether this becomes an issue for Airbnb as the number of ratings it has in the North American cities surveyed in November and May seems to have peaked. How will it continue the growth that is demanded of it by its backers in the face of the large number of hosts who drop off the platform?

While much is made of the novel peer-to-peer nature of Airbnb’s reputation system, it appears that the company runs a complaint-based trust system that is not so different from other hospitality companies. In April Airbnb expelled 2,000 listings from its New York offerings, claiming that they did not offer a positive experience, but many of those listings had either no ratings at all or were rated highly (4.5 or 5) by their guests: it appears that Airbnb did not use its own ratings when judging who to remove from its platform.

Uber Drivers Earning $90K/year? More Evidence Needed

Last Tuesday taxi-disrupting tech company Uber posted on the company blog that “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”. I don’t think their numbers add up, but first the story so far…

For the notoriously cheap taxi industry, those are some pretty sweet numbers, and they were hailed (hah!) by Matt McFarland of the Washington Post with the headline Uber’s remarkable growth could end the era of poorly paid cab drivers. Noting that “estimates of the typical cab driver’s salary hover around $30,000”, McFarland acknowledged that “Uber’s numbers don’t account for the costs a driver incurs to own and operate a vehicle. Still, the gap in compensation for providing similar services is astounding”.

As the story spread, many just ignored those pesky costs. CNBC led with “Uber’s $90K salary could disrupt the taxi business”. The New Orleans Times-Picayune headlined its story “Uber drivers in New York City earn more than $90,000 a year, newspaper reports” and Entrepreneur.com claimed “The Median Income of an Uber Driver in NYC Is Nearly $100,000”. From the technology industry, CEO Mike Jones laid it out at Code Conference: “You’re qualified to drive a car, but not professionally doing it. Congratulations, boom, you’re making [a] $90,000-a-year average Uber salary.”

Among all this enthusiasm, a few voices did raise some questions. In Time Magazine Dan Kedmey emphasized the costs:

Unfortunately, the figure excludes many of the costs of running a business, including gas, insurance, parking, maintenance and repairs and the original sale or lease price of the car which can take some hefty bites out of the driver’s take home pay. It also measures a median income among a particularly dedicated set of drivers, logging a minimum of 40 hours a week and sometimes much longer hauls.

He asked Uber, but they shrugged their shoulders.

Just how much those costs eat away at a driver’s take home wages is not easily gauged, according to Uber spokesman, Lane Kasselman. They can vary depending on the age of vehicle, the density of app users in the city, how many hours the driver puts in and what sort of customer ratings the driver receives. And for now that data, Kasselman says, is proprietary.

At Mashable, Jason Abbruzzese asked whether the income is sustainable (Uber is looking to entice drivers and has deep pockets, after all), and at The Atlantic’s CityLab Eric Jaffe pointed to the Uber driver protest in San Francisco, where drivers claimed they “work for less than minimum wage” and asked how these stories could fit together.

Back in December, economists Felix Salmon (here) and Tim Worstall (of Forbes) had both had fingered the culprit for taxi drivers’ appalling incomes: the medallion system that many cities use means that medallion owners get to take the money. Now, in the wake of the new claims, Salmon asked Uber for details of these extra expenses and they actually sent him numbers (take that, Time Magazine!) showing that business expenses would be around $15K, so the drivers are still making twice the norm: not bad. Salmon deemed these numbers “reasonable and entirely intuitive”.

Uber, by the way, takes 20% of the fare, plus a $1 “safety fee”.

Now I’m no expert, but I thought I’d take a look at some reports into the taxi industry and see what I could find out about the Uber claims. The short version is this:

  • The medallion owners in some cities take roughly the same amount of the fare as Uber. They may be ripping off the drivers, but the lease costs that they charge don’t seem to be the main reason for the difference between the numbers. I think Salmon and Worstall have this wrong.
  • The Uber claims over car maintenance costs are under what other reports say about the costs. I suspect they are cherry-picked, but they are also not the main reason for the difference.
  • The big difference comes from the claim that an uberX car takes in $110,000 in fares over a year, while driving 40,000 miles. A regular taxi takes in about half that, and drives about 50% more. It’s not clear where this difference comes from (different cities? different ways of counting? bad guesses?) but if Uber is going to stick by its claim, it needs to explain the difference.
  • The media writers who take business expenses as a minor factor in the driver’s overall income are way off.
  • Put this all together, and the driver incomes look too high.

First let’s look at the plight of taxi drivers. I found relatively recent reports on the taxi industry in three major North American cities: a UCLA study on Los Angeles (2006), a San Diego State University report on San Diego (2012), and two reports about Toronto (2008, 2012). Obviously these are not San Francisco and New York, which is what Uber was writing about, but the point is not to ask if they are telling the exact truth, but to see if the picture they paint is a representative one.

The three reports paint a grim picture of a taxi driver’s life. In all three cities, most drivers work six 12-hour shifts a week for less than minimum wage, even after tips. These are mainly immigrant men, and most are between 30 and 50 so many have family responsibilities. Health insurance is non-existent, and the job is dangerous with assault, vomit (which they have to pay to get cleaned) and petty aggravations as perpetual companions. I don’t think anyone wants to paint a rosy picture of the taxi driver’s working conditions.

But now to the numbers. There is a surprising consistency to the three cities (which I write as LA, SD, and TO).

  • In their 72 hours of driving each week, a driver will cover about 1200 miles (LA), of which just under half (LA) are “on the meter”. The LA study gives an average of 6.2 miles-per-paid-gallon.
  • The total income (fares + tips) that comes into the cab over a week is about $1500 (LA), $1100 (SD), or $1150 (TO).
  • Gas is a significant expense, and cab drivers have to pay it. The weekly cost is about $250 (LA), $260 (SD), and $250 (TO). This suggests that San Diego drivers cover the same distance as LA, while Toronto (where gas is more expensive) cover significantly less.
  • Most taxi drivers pay a lease that covers car insurance and maintenance as well as earning money for the medallion owner. The weekly lease is about $500 (LA), $400 (SD), and $260 (TO). There’s quite a bit of variation in each city because there is a mixture of owner-operators, shift drivers, and others at work and their situations are all different.
  • The Toronto report separates the car maintenance, depreciation, repair and insurance out. For LA and SD this will come out of the medallion owner’s income. It is about $300/week ($70 insurance, $70 maintenance and repairs, $175 on car financing).
  • That leave a weekly income for the driver of about $600 (LA), $320 (SD), and $450 (TO). This comes to an annual income of $31,000 (LA), and a terrible $16,60 (SD) and $23,400 (TO).

We can put these in a table and compare them to the Uber estimates:

[Update: changed these figures on June 3 to be better averages of the driver kinds in each city.]

Quantity (weekly) LA SD TO Uber (est)
Fares + tips 1500 1100 1150 2070
Gas 250 260 250 115
Lease/Operator 250 140 260 414
Car Operation & Depreciation 300* 300* 300 76
Other expenses 100 70 90 0
Driver Income 600 330 250 1465

* = uses Toronto estimate, as no estimate given in report. Is paid by the lease-holder for those who lease.

Do the Uber numbers make sense? At the very least they need some explaining before we take them seriously. Here are the questions that need answering:

  • According to the Washington Post, Uber’s sample is “drivers working over 40 hours per week”. Are they working the same 72 hour weeks that taxi drivers are?
  • Taxi mileage is way higher than Uber’ estimate of 770 miles per week. Maybe Uber is not counting the “off meter” miles?
  • Can they justify the low gas costs that they estimate or are they not counting time between rides (which is half the mileage for taxi drivers)?
  • Uber seems to take more than the leaseholder (after expenses). Given the slagging off that medallion holders get, this surprised me.
  • The car maintenance fee for Uber is much lower than for taxi drivers. Does this reflect a lower standard for Uber or where do they get this number?

In short, a lot of questions to be answered before Uber’s claims can be justified.

The shape of Airbnb’s business

When Airbnb talks about its legal troubles in New York, Berlin, Amsterdam, and elsewhere, it claims that existing laws were never designed for its new brand of disruptive peer-to-peer business.

There were laws created for businesses, and there were laws for people. What the sharing economy did was create a third category: people as businesses… They don’t know whether to bucket our activity as person or a business.

In 2010, the State of New York passed a law designed to crack down on bad actors that operate illegal hotels—a goal we all share. Unfortunately, the 2010 law also had the unintended consequence of impacting regular New Yorkers.

There are more. You get the point.

In a series of studies designed to address regulators’ concerns, Airbnb talks about its hosts as “regular people” and focuses on the ways its business is different to the existing tourist business. It highlights the hotel industry as a point of comparison and emphasizes just how different Airbnb is. Here are a few typical quotations:

  • “87 percent of hosts rent the homes they live in” (Amsterdam)
  • “87 percent of Airbnb hosts rent out the home they live in, and the typical host earns $7,530 per year” (in New York)
  • “Airbnb is complementary to the existing tourism industry in Paris. 70 percent of Airbnb properties in Paris are located outside the central hotel corridor.”
  • “73 percent of Airbnb properties in Amsterdam are located outside the eight central tourist districts.”
  • “Airbnb’s 5,600 local hosts are regular people who occasionally rent out their homes and use the income they earn to pay the bills.” (in Berlin)
  • “About 80% of Airbnb hosts rent out the home they live in” (in London and Edinburgh)

There are more. You get the point.

The claim of novelty and of hosts as “regular people” has been widely accepted. For example, the thoughtful Kevin Roose wrote this in New York magazine the other day:

There are no laws governing Airbnb because until very recently, there was nothing like Airbnb in the world—not of the same scale, not with the same guiding philosophy. And when Airbnb came onto the scene, regulators were forced to slot it into existing categories where it, arguably, didn’t belong—treating a bachelor renting out his spare room to make rent, for example, with the same rules as a scuzzy landlord operating an illegal hotel. They’ve been playing catch-up ever since.

Or as Wired Magazine writes:

We are hopping into strangers’ cars (Lyft, Sidecar, Uber), welcoming them into our spare rooms (Airbnb), dropping our dogs off at their houses (DogVacay, Rover), and eating food in their dining rooms (Feastly).

There are more. You get the point.

So when it comes to thinking about and dealing with this and other sharing-economy companies, the kind of business that Airbnb operates matters. Does it match the company’s self-portrait? Is the company as novel as it claims to be? Are its hosts “regular people”?

It turns out there is an element of wishful thinking in the portraits of Airbnb. Last November, I took a look at Airbnb data from New York (here and here). In February, travel web site Skift carried out a similar study (here and here), which was part of the New York Attorney General’s case against Airbnb. The two studies took similar approaches, collecting listings from Airbnb’s public web site and gleaning what we could from that imperfect data set. Both studies concluded that, while it is true that a large number of hosts rent the homes they live in, hosts with multiple listings make up almost half of Airbnb’s business. Also that, while Airbnb makes great play of its origins in renting out an airbed, such rentals are now a negligible portion of its business. Even “spare rooms” are a minority of the business: the majority of Airbnb’s business in New York comes from the rental of entire homes.

The data showed a company that was closer to orthodox models such as HomeAway and its subsidiary VRBO than the narrative would have it. There are differences—HomeAway is focused on vacation rentals, and many of its properties are run by property managers—but the similarities cast doubt on Airbnb’s claims that existing regulations are inapplicable.

Now here we are: it’s six months on, and interest in Airbnb continues. Airbnb has kicked 2,000 New York listings off its site (10% of the total for the city). It handed over host data to the Attorney General (anonymized, the company says). Meanwhile, the company is valued at $10 billion, having raised $450 million in a new round of venture capital. The New York dispute is now over, but the sharing economy poster child is still here, bigger than ever, and still a leading light in the wave of digital disruptors looking to shake things up and make a lot of money.

So during May I collected data on over 90,000 hosts and 125,000 listings—about 20% of Airbnb’s 600,000 total, according to this TechCrunch estimate—from 18 cities around the world, to sketch a portrait of Airbnb’s business. The main questions I had in mind are the straightforward ones, starting with the same ones Skift and I asked about New York:

  • Is Airbnb’s business based on “regular people” in a way that other part of the hospitality industry are not?
  • Is Airbnb’s business based on spare rooms and airbeds?
  • I looked again at several cities I had collected (but not posted about) in November, so that I could look at how the business has changed in some of Airbnb’s key markets.
  • I hoped that looking a second time at New York might have something to say about the 2,000 listings that Airbnb removed from the site in April, during its run-in with the Attorney General.

For those who don’t want to read the whole thing, here are the quick answers.

  • While a good part of Airbnb’s business is based on “regular people”, over 40% comes from hosts with multiple listings. This is different from Airbnb’s self-portrait. Airbnb’s claim that existing regulations don’t apply to it is at least exaggerated.
  • The majority of Airbnb’s revenue comes from whole-home rentals. This makes the company much more like HomeAway and other vacation rental businesses. It casts further doubt on the company’s claim to be a new class of business.
  • In some of its biggest markets, Airbnb may have maxed out the number of listings it can achieve. What’s more, there is a high rate of churn as individual hosts put a property on the market, have a few guests, and then take the property off again.
  • Airbnb does not appear to believe its own claim that customer ratings provide an assurance of good experience. Airbnb says it removed 2,000 New York listings from the site because of bad experience, but at least half of those listings had good (4.5 or 5 star) average ratings from customers.

I’m going to post this in two parts.

  • Part 1 looks at the split between hosts with a single listing and those with multiple listings, and it also looks at how far Airbnb has moved from its “origin myth”, after which the company is named—the hosting of people on couches and in shared rooms.
  • Part II looks at the change in Airbnb’s business over the last six months, including the changes in New York where the legal strife has been the loudest.
  • If I have time and if there are requests, I’ll collect these three together and post a single PDF.

Background and Method

Just to set out some basic information, Figure 1 shows the number of listings and of hosts in each city. The data was collected using a fairly straightforward two-stage search. The first stage goes through all the search pages for a specified city and collects the room_id values. The second stage visits the room page for each value and gets details about the listing. The code is available on github.

A few notes:

  • Airbnb has regularly said that is has about 20,000 listings in New York city. I find 19094 (and have over 20,000 from November) so the collection seems pretty complete.
  • In 2013, Airbnb claimed 5600 hosts in Berlin. I find 6141.
  • In the run-up to the World Cup, Airbnb claims to have 9,000 listings in Rio. I find just over 10,000. Perhaps the borders of the search are different, or the number is growing. Still, 90% accuracy is not bad.

In short, the surveys for each city seem pretty accurate.

city_basics.png

Figure 1: Number of hosts and listings in the surveyed cities

The host perspective

Figure 2 shows the percentage of hosts in each city that have a single listing on the site. The values for New York and Amsterdam match Airbnb’s claim for New York and Amsterdam to within a percentage point, which suggests that the sample is realistic and that the use of a single listing is a pretty good proxy for “regular people who occasionally rent out the home in which they live”. The graph also shows that the claim applies to most of the big cities. Barcelona, Rome, and Tokyo are the only Airbnb locations surveyed that have fewer than 80% of hosts with a single listing. So far, so good for Airbnb’s self-portrait. From here on I will call hosts who have a single listing “regular people”.

city_hosts.png

Figure 2: Percent of hosts with a single listing

The marketplace perspective

Imagine that a city has 100 hosts, 99 are “regular people” with a single listing and one host has 99 listings, then the percent of hosts who are regular people would be 99%, but the percent of listings on the market that come from regular hosts would be only 50%. Both percentages are important in gauging the kind of business that Airbnb is. Figure 3 shows the percent of listings offered by regular people. The overall figure is 62%: still a significant majority, but a number that is 20% lower than the percentage of hosts.

We can see that for a few cities, notably Barcelona (11,000 listings) and Rome (growing quickly, at 8,000 listings), the majority of listings come from hosts with more than one offering.

city_listings.png

Figure 3: Percent of listings from hosts with a single listing

The traveller perspective

The percentage of listings that come from different types of hosts corresponds to the experience of the potential guest browsing or searching the Airbnb site. The traffic generated by Airbnb is different, because not all listings are equally popular. There may be areas with many Airbnb listings but relatively few actual visits, while other areas may have listings that are visited very frequently. Airbnb does not give the number of actual bookings for each listing (or, nearly equivalently, the number of visits to the listing), but it does give the number of reviews that each listing has received, and this should be a reasonable proxy for the number of visits.

Over a third of all listings have no reviews at all, and some have many (the most-reviewed listing in my sample is this San Francisco treehouse, a novel place to stay with 460 reviews.)

Figure 4 shows the percentage of visits that are to rooms listed by regular people. The numbers are getting significantly smaller now. No city has more than three quarters of its visits at properties of regular people, and several have a slim majority of visits to hosts with multiple listings. Overall, 45% of visits happen at places offered by hosts with multiple listings.

city_bookings.png

Figure 4: Percentage of visits (bookings) to listings offered by hosts with a single listing, using reviews as a proxy for visits

The Airbnb perspective

There is one more step to take, which is to look at Airbnb’s actual revenue. What fraction of its business comes from regular people and what fraction comes from multiple listers?

To get here, I multiply the bookings (proxied by reviews) with the listed nightly price. Again, it’s not a perfect measure but so long as regular hosts don’t overall have longer- or shorter-stay guests compared to other hosts, it should give a reasonable picture.

Figure 5 shows the percentage of Airbnb’s revenue that comes from regular cities. Only three cities have over 60% of their revenue coming from hosts with a single listing. Overall, 44% of Airbnb’s business comes from hosts with more than one listing, which is slightly up from 42% in November (the increase may not be significant).

city_revenue.png

Figure 5: Estimated percentage of Airbnb revenue from hosts with single-listings

Airbnb listing types

The story of Airbnb emphasizes the casual “airbed” rental, but this is a very small part of Airbnb’s business. Again, there are a couple of ways of looking at the data.

Listings by room type

Every listing on the Airbnb site is listed as one of three categories: a private room, a shared room, or an entire home/apartment. Figure 6 shows the breakdown in each city. It is clear that shared rooms are a negligible portion of the total: about 1 in 50 listings are shared rooms.

city_listing_roomtype.png

Figure 6: Number of listings that are private rooms, entire homes or shared rooms

Visits by room type

Figure 7 shows the visits by room type, using reviews as a proxy for visits again. Shared rooms are an even smaller percentage of the whole: only 1.4% of Airbnb visits are made to shared rooms.

city_booking_roomtype.png

Figure 7: Number of visits to private rooms, entire homes or shared rooms

Revenue by room type

Using reviews * price as a proxy, Figure 8 shows the percentage of revenue from each room type. The revenue that comes from spare couches and shared rooms is a mere 0.56%.

city_revenue_roomtype.png

Figure 8: Revenue from different types.

So what?

  • The data show that Airbnb is consistently economical with the truth when it describes its own business. It’s a long way from being just “regular people” and there is a lot more business coming from multiple-listing owners than they let on.
  • The data say that it’s time commentators and the media stopped using the”Couchsurfing” narrative for Airbnb. At less than 1% of its business, the couchsurfing model is irrelevant for what Airbnb is today. It’s far more like HomeAway than it is like Couchsurfing.
  • If there is a novelty to Airbnb’s business, it’s that it collects property managers, individual renters, and occasional renters under one roof, just as Amazon can offer best sellers, midlist authors and self-published obscura in the same place. And while much of the talk will be about the long tail of rooms, the reality is that Airbnb is pushing for the professionalization of its hosts, and we’ll see more of that over time.
  • Other conclusions to come after we see the rest of the data.

Out and About

Despite the slack posting here (but I do have some things I’m working on, honest), I am flattered that people still want to talk to me about sharing economy and related topics. Here are some recent posts and articles where I appear.

  • I’m going to be talking sharing with a panel of people at Personal Democracy Forum 2014. Unusually for a technology-focused conference, over half the speakers are female, which is great. The blurb for my pane; says “Defining and Debating the “Sharing Economy” (Thurs June 5, 3:30-4:30pm): Tom Slee, James Slezak, Denise Cheng, Adam Greenfield, Nancy Scola (moderator). Is there such a thing as the “sharing economy,” or are some people more interested in sharing, while others are interested in the economy? How can we usefully distinguish between “sharing economy” enterprises that empower their users and those that may actually seek to exploit them? A critic [that’s me], an advocate and an academic observer will take a close look at these questions and try to get to the bottom of what is good and what is hype about this emerging phenomenon.”
  • Terry Dawes of CanTech Letter is right. Jeremy Rifkin’s views on the future are indeed, neat, simple, and wrong.
  • Jon Zerolnik of LA-based investigative website Capital and Main interviews me about Couchsurfers and Billionaires.
  • Thanks also to Houston Davidson of the College Hill Independent for talking to me about Sharing for Money.
  • The always-excellent Nancy Scola writes about Arun Sundararajan and quotes me.
  • Hey, look! My non-earth-shattering 1988 paper with Preston McDougall on “The correspondence between Hückel theory and ab initio atomic charges in allyl ions” is now online!

The People’s Platform: Who are No Logo’s Children?

Of all the books I’ve read about digital technology and its effects on our culture, Astra Taylor’s The People’s Platform is closest to my own beliefs. I think it’s a wonderful book, but what I really want is to see more people disagree with it.

The trouble is, everyone seems to like The People’s Platform. In Canada, The National Post likes it and The Globe & Mail calls it “a No Logo for digital natives”. In the UK The Guardian says it is an “invaluable primer” for understanding our networked world, The Telegraph gives it four stars out of five, and The Financial Times is positive. In the US I haven’t seen reviews in the major papers except for The Boston Globe, which is persuaded by Taylor’s arguments; but on the left, Dissent and Tom Dispatch welcome it, while Kirkus Reviews calls it “a cogent and genuine argument for the true democratization of online culture”.

The thing is, the book is a challenge to those who see themselves as digital progressives, or part of a digital counterculture—the descendants of No Logo, perhaps, who would trace a disruptive, counter-cultural path from the anti-copyright campaigns of the 1990s through Free/Open Source Software to net neutrality campaigns, Creative Commons, Open Data, Pirate Parties, the Arab Spring, Occupy, and Anonymous.

Not that we need to lump these all together of course, but there is a digital-progressive belief that the design of the Internet gives it a unique potential among technological innovations to be democratizing and liberating—and that (despite disenchantment with the advertising giants) this potential has been validated by the ways in which people have used digital technologies to challenge existing power structures.

It’s a “Sympathy for the Digital”, if you like: a willingness to give digital disruptions the benefit of the political doubt—to overlook the Wall Street connections of Bitcoin, or the ways in which open source institutions have aided the NSA’s spying activities—and a readiness to indulge today’s billionaires when they adopt the anti-corporate, counter-cultural language of the 1990s. It’s that feeling that the Pierre Omidyars of the world are somehow on “our” side because they come from the technology world. It’s the reason why the Electronic Frontier Foundation intervened on behalf of Airbnb against the State of New York. It’s where left-leaning people meet libertarians.

Taylor’s sympathy for the digital has run out. It ran out, perhaps, when the documentary she spent two years of her life making was loaded up onto download sites and her requests for a short grace period were rejected because “philosophy is free”. The preciseness of our vocabulary tells us how we look at the world, and reducing cultural creativity to what Taylor calls “that flattening word, content” is a condemnation of the digital world’s indifference to artistic work.

The way she sees today’s digital landscape, “we are witnessing not a levelling of the cultural playing field, but a rearrangement… In the place of Hollywood moguls, for example, we now have Silicon Valley tycoons (or, more precisely, we have Hollywood moguls and Silicon Valley tycoons”. As we look at how the last 15 years have turned out, the big picture that Taylor paints is not one of egalitarian progress, but of a sort of Animal Network, in which the new rulers have taken on the manners and values of the old, while clinging to the rhetoric of rebellion and change to justify their actions.

She takes on a wide range of topics, including equality, openness, the challenges of sustaining creative labour, investigative reporting as a public good, the winner-take-all nature of Web 2.0 platforms, the value of limited copyright, the environmental impact of online culture, gender imbalance in the technology world, the implications of an advertising (and self-promotion) driven culture. She does so with a wide range of reading, and in an accessible style, in a voice that is intelligent and full of conviction. The book is not an academic book or a “big idea” book structured around a populist, easy-to-digest nugget. It’s an essay, in the best sense of that word.

Her broad conclusion is that the digital world has not turned out the way that the enthusiasts and evangelists of the last two decades have claimed it would. In addition to predictable targets like Kevin Kelly, Jeff Jarvis, Don Tapscott and the more libertarian strands of Wired culture, Taylor takes on leading digital progressives such as Lawrence Lessig, Clay Shirky, Yochai Benkler, Steven Johnson, and Tim O’Reilly: people who would think of themselves as egalitarian. She places them, broadly, on the wrong side of current debates about the impact of new technologies on culture and society. The book suggests that it’s time these digital progressives engaged in some serious questioning of their own viewpoints; that they acknowledge that we live in different times now, and that the ideas and arguments of 1994, or even 2004, have different implications in 2014. I don’t know if Taylor would do this, but I’d extend her argument to people like Biella Coleman, whose “Coding Freedom” is a smart example of digital counter-culture thinking, and to the kind of thinking that comes out of the Berkman Center.

Taylor’s book is provocative partly because of her own background. As an independent film maker and Occupy activist, she has a history of frustration with mainstream media; in her own words, she is “a prime candidate… for cheering on the revolution that is purportedly being ushered in by the Internet” (p2). We all cluster in camps of one kind or another, and for digital progressives it is easy to argue that a book like Robert Levine’s /Free Ride: How Digital Parasites Are Destroying the Culture Business, and How the Culture Business Can Fight Back/ comes from a place of entrenched interest and reaction. It is more difficult to paint Taylor as a digital conservative: this is criticism from inside the tent.

This is why I’m so disappointed that there has been no serious defence of digital progressivism in response to The People’s Platform. (Have I missed some? Please leave links if so.) The book deserves to spark a debate, but it takes two to tango, and the digital progressives seem to be sitting this dance out. (I would not expect the Kevin Kellys of the world to respond, but I had higher hopes of others).

Why the lack of retort? Could it be that Taylor does not have the profile to justify a response? That it is easier to ignore The People’s Platform and wait for it to go away? It certainly looks that way from here. Let’s hope I’m proved wrong.

In the absence of a strong response, it’s time to say that digital progressivism has no legitimate claim to the political or counter-cultural legacy of No Logo and the movement that book embodied. Sure, part of that political legacy was the anti-copyright campaigns of the early 2000s. Sure, digital initiatives like Indymedia came out of that movement. But that movement was also a protest against branding, and against the McDonaldization of global culture and the American cultural imperialism that accompanied it. For those of us outside the US, the right of individual countries to set their own rules to sustain their cultural industries and their cultures were part of the struggle. If that’s a struggle that runs counter to what some call the logic of the Internet, well so be it.

No True Airbnb Host…

After months of waiting, Airbnb and the Attorney General of New York finally face each other in court this week to argue over the Attorney General’s subpoena for information on the 15,000 or so Airbnb hosts in New York City (summary at the Guardian, more opinion in the overview by Nitasha Tiku at Valleywag). There’s been a flurry of revealing activity and statements, and lots of commentary. So here’s more.

The court case is about many things. But what makes it interesting for those of us who don’t live in New York is the broader implications of the case, which include:

  • The breadth of the Attorney General’s subpoena and the proportionality of his actions. Does the subpoena blend into the general fear of government access to individual data in the light of the NSA scandals?
  • The illegality of some Airbnb rentals. If the subpoena succeeds and shows that a significant portion of Airbnb’s business is based on illegal rentals, then that’s a problem for a lot of people.
  • If Airbnb rentals are illegal, who is responsible (legally and morally)? Is it the host or Airbnb itself?
  • The validity of the law. Airbnb maintain the law is a bad law, inappropriately applied, and that the Airbnb hosts are a new class of business (“micro-entrepreneurs”) that need new rules. Does new technology render the old laws obsolete?
  • How will this affect the $10B valuation of Airbnb and its prospects for an IPO?
  • How will this affect the future of the sharing economy?

The main players are, obviously, Attorney General Eric Schneiderman and Airbnb. But other players appear on the stage, including:

  • Airbnb hosts and guests. Hosts have the most to lose if their information is handed over.
  • Democrat Senator Elizabeth Krueger. Her position on the issue (see here and today’s interview by Nancy Scola) is a bit different from that of the Attorney General, but she is broadly on the same side.
  • The technology industries and advocates. The Internet Association and the Electronic Frontier Foundation have both spoken up about the subpoena, placing themselves broadly on the Airbnb side.
  • Neighbours, landlords, tenants, and housing co-ops. These are all affected by individuals renting out apartments through Airbnb.
  • The hotel industry. While Airbnb has painted hotels as one of the villains in the fight, their role seems to me peripheral.

To start with the narrowest issue: it seems pretty clear that about 2/3 of Airbnb rentals in the city are outside the law in New York, which forbids apartment rentals of under 30 days if the owner is not present. The New York Post report is here, and is based on an affidavit by the Attorney General’s office (PDF). That affidavit is, in turn, based on an analysis of Airbnb’s business carried out by consultants Connotate for travel site Skift. Their report is more extensive than, but similar to, the analysis I carried out a while ago here. Airbnb and Peers recognize this by arguing that the law needs to be changed.

For what it’s worth, I find Elizabeth Krueger’s position on illegal rentals more reasonable than either the Attorney General’s or that of advocacy group Peers. The Attorney General gives the impression of (at least potentially, and despite avowals otherwise) going after all the hosts who have made illegal rentals. Given that law enforcement is usually complaint-based and that Airbnb has done little to warn its hosts of their responsibilities, this seems harsh. Krueger (see links above) puts Airbnb as the root of the problem—making a lot of money off hosts and letting the hosts take all the risk—and I agree with her. Airbnb may not be legally responsible, but it seems to me morally culpable.

But what if the law is just a bad, obsolete law that shouldn’t apply to Airbnb hosts anyway, which is what Airbnb has been arguing? Here the onus switches to Airbnb: has it made the case that it can do better than New York’s “bad law”? It needs to do more than say “because Technology” if it’s going to justify a change.

Most of the Airbnb case is made in its public policy blog posts (here and here), in its report on New York (here), and in its new sharing cities initiative (here). What these posts show is a remarkable lack of content, and a reliance on heartwarming words and spin that is, in the end, cheap talk.

Airbnb’s description of its own hosts is at least consistent. They repeatedly describe their “community” as “regular New Yorkers who occasionally rent out their homes”, or “regular New Yorkers just trying to make ends meet”. They emphasize that “87 percent of Airbnb hosts rent out the home they live in”. These are all phrases that invoke the “sharing economy” vision. When anything goes wrong, Airbnb refers to occasional, incredibly rare “bad apples”.

But Airbnb’s homespun language is carefully chosen. The 87% figure is the most obviously economical with the truth. It may be true as a percentage of hosts, but only about 63% of listings are single-listing hosts, and as much as anyone can tell, almost half of actual bookings are from hosts who have multiple listings. It’s not easy to get definitive answers from scraping the public web site, but Airbnb has consistently refused to challenge these numbers, and so it’s a good guess that they are not too far off the truth. They could release their own statistics if they chose, but they don’t.

Instead, they put out “reports” that have no methodology, no definitions of key terms, and about 300 words of actual text. Airbnb makes a lot of the fact that it “supports more than 4,500 jobs” but it never says what “support” means. It claims to have “generated $104 million in economic activity outside of Manhattan”, but it doesn’t say what “generated” means. This may seem like nit-picking, but it isn’t: it’s just asking for facts instead of spin.

But where things get really bad — for Airbnb hosts who drive the company’s revenue as well as for Airbnb’s claim to be a responsible provider of safe and well-managed accommodation — is in Airbnb’s slippery definition of its own “community”.

Airbnb continually claims to speak for its hosts, as in “our hosts want to pay taxes”, and repeatedly characterizes its community of “amazing” hosts as “regular” people. It also describes them as “micro-entrepreneurs”, suggesting that they are independent. But Airbnb hosts, who are asked increasingly to invest in their own property in order to put it on Airbnb’s marketplace, are precariously dependent on Airbnb’s whims in order to make money from their investment, which is not an entrepreneurial role at all. If anything goes wrong, the host gets turfed off the market without appeal. If you’re a host, how do you know if you are a “bad apple” or a “regular New Yorker”?

Most dramatically, this has happened today. After insisting for months that concerns about multiple listings were exaggerated, Airbnb today permanently removed no fewer than 2,000 New York listings from its platform – roughly 10% of the total. They claim the process has been going on for months, and that it demonstrates their responsibility. But the fact that it happened today of all days, the fact that identifying multiple listings is a trivial exercise, and the fact that the criteria for expulsion are still woefully unclear, makes it difficult to take them seriously (Jason Clampet at Skift is excellent on the removals). Here is an excerpt of Airbnb’s explanation:

But when we examined our community in New York, we found that some property managers weren’t providing a quality, local experience to guests. These hosts weren’t making their neighborhood stronger and they weren’t delivering the kind of hospitality our guests expect and deserve. In some cases, they were making communities worse, not better. We took a hard look at our community in New York to identify these hosts and we took action.

Earlier this year, we began notifying these hosts that they and their more than 2,000 listings would be permanently removed from the Airbnb community. While we are allowing these hosts to support their existing  bookings, all are now prohibited from accepting new reservations and if you search for a place to stay in New York, you won’t find these listings.

Imagine you are an Airbnb host in New York reading these words. How would you know if you are living up to Airbnb’s requirements? Are you making your neighbourhood stronger? Are you delivering the kind of hospitality your guests expect? This is a “No True Scotsman” argument: the Airbnb community is trustworthy because if you do anything Airbnb decides is wrong then you are not part of the community. The lack of clarity is remarkable and shows that they are not ready to provide the kind of security and accountability that any replacement for rentals regulation would need.

My skepticism over Airbnb’s sincerity is heightened because, apparently, one blocking point in the negotiations with New York is that (according to Clampet), “Airbnb would not agree to limits on how many listings a person could have in New York City.” It’s clear that a large number of bookings is essential to the venture capital model, and goes against the spirit of the “sharing economy” that Airbnb so consistently invokes. But it looks like they’re going where the money is. That makes them untrustworthy.

Airbnb’s evocative but meaningless cheap talk (for more, see their Shared Cities initiative) is cynical. I fear that, instead of promoting any realistic idea of sharing, Airbnb will pollute the whole idea as a consequence of the high-return venture capital model it has pursued.