ArticlePDF Available


Airbnb is an Internet-based firm that connects potential short-term renters with hosts who own or control rental properties. Its rapidly expanding activities are tracked by Airdna, an independent firm that generates seemingly conventional performance metrics describing Airbnb. These metrics include occupancy rates, average daily rates, and revenue per available room. However, Airdna does not adhere to long-established STR definitions for these variables. Using data from Virginia Beach, Virginia, we demonstrate that Airdna’s performance metrics exhibit notable upward biases vis-á-vis STR’s metrics. Potential rental hosts, hoteliers, tax collectors, and investors are at risk if they act on the assumption that Airdna’s metrics are comparable with widely understood measures used by STR and tourism experts.
Cornell Hospitality Quarterly
2019, Vol. 60(3) 193 –199
© The Author(s) 2018
Article reuse guidelines:
DOI: 10.1177/1938965518777218
The increasing success of Airbnb, an international, Internet-
based firm that connects owners of rental properties with
short-term renters, has attracted increasing attention since
its inception in 2008. Airbnb is now active in more than
65,000 cities and 191 countries and has facilitated more
than 160 million stays (Airbnb, 2017). By mid-2017, Airbnb
claimed to have more than 3 million rental listings, raising
concerns of disruption of the traditional lodging sector and,
more locally, negative externalities such as loss of tax rev-
enue, traffic congestion, noise, and disruptive behavior
(Edelman & Geradin, 2016; Guttentag, 2015). Proponents,
on the other hand, note the increased income accruing to
property owners, improvements in economic efficiency,
and the emergence of the “gig” economy for cities and
municipalities (Koopman, Mitchell, & Thierer, 2015;
Quattrone, Proserpio, Quercia, Capra, & Musolesi, 2016).
Although Airbnb recently launched a public Application
Programming Interface (API), this API is not sufficiently
open for researchers and firms to access individual listing
data across the Airbnb platform.1 Hence, even though
Airbnb’s impact is of great interest to many different indi-
viduals, assessing that impact may require the use of data
that present a skewed picture of Airbnb’s activities.
Finding useful data on Airbnb and similar entities is
increasingly important given the significant challenges these
firms pose to the operations of the traditional lodging sector.
Home-rental market revenue is only about one fifth the size
of the traditional lodging sector but is growing at a faster
pace than the hotel industry (Kirkham, 2017). Emerging evi-
dence suggests that Airbnb and similar products not only
expand the supply of available rooms but also reduce con-
ventional hotel revenue on high demand dates by up to 40%
(Jordan, 2015). The greater the market penetration of Airbnb,
the lower the average price for conventional hotel rooms
(Neeser, Peitz, & Stuhler, 2015). The economic impact of
Airbnb is not uniform, however, with lower-priced hotels
and those catering to nonbusiness travelers being most
affected over time (Zervas, Proserpio, & Byers, 2017).
Recent survey evidence of Airbnb users indicates that nearly
two thirds used Airbnb as a hotel substitute (Guttentag &
Smith, 2017).
Similar evidence suggests that Airbnb benefits the tour-
ism industry by lowering tourism costs, leading to increased
spending on nonlodging services and goods. Lower-end
hotels, however, are displaced by Airbnb, suggesting a trad-
eoff between gains for the nonlodging sector and the lower
end of the lodging sector (Fang, Ye, & Law, 2016). Finding
useful data with regard to Airbnb rentals, however, is often
an arduous task in that researchers must either manually
sample the data (which is likely to introduce errors), scrape
the data from Airbnb (which is often not replicable as the
777218CQXXXX10.1177/1938965518777218Cornell Hospitality QuarterlyAgarwal et al.
1Old Dominion University, Norfolk, VA, USA
Corresponding Author:
Vinod Agarwal, Department of Economics, Strome College of Business,
Old Dominion University, Norfolk, VA 23529, USA.
Differing Views of Lodging Reality: Airdna,
STR, and Airbnb
Vinod Agarwal1, James V. Koch1, and Robert M. McNab1
Airbnb is an Internet-based firm that connects potential short-term renters with hosts who own or control rental
properties. Its rapidly expanding activities are tracked by Airdna, an independent firm that generates seemingly conventional
performance metrics describing Airbnb. These metrics include occupancy rates, average daily rates, and revenue per
available room. However, Airdna does not adhere to long-established STR definitions for these variables. Using data from
Virginia Beach, Virginia, we demonstrate that Airdna’s performance metrics exhibit notable upward biases vis-á-vis STR’s
metrics. Potential rental hosts, hoteliers, tax collectors, and investors are at risk if they act on the assumption that Airdna’s
metrics are comparable with widely understood measures used by STR and tourism experts.
Airbnb, Airdna, occupancy rate, average daily rate, revenue per available room, Smith Travel Research, hotel industry,
rental properties
194 Cornell Hospitality Quarterly 60(3)
information on the site changes daily), or rely on others to
produce information on Airbnb hosts and guests.
Airdna is a firm that generates data and analytics focus-
ing on short-term rental entrepreneurs and investors. Airdna
sells Airbnb performance data and is the dominant supplier
of data concerning Airbnb’s performance. Airdna’s unique
position is based in part on a proprietary algorithm that dis-
cerns whether specific Airbnb rental properties are avail-
able, reserved for occupancy, or blocked by the owner and
unavailable for rental.2 A large (and growing) number of
public and private decision makers utilize Airdna’s data to
explore the impact of Airbnb on the traditional lodging sec-
tor, renter behavior, rental pricing, and investment profit-
ability. Comparing Airdna data with seemingly comparable
data from STR (formerly Smith Travel Research) concern-
ing the performance of the traditional lodging sector is not
only a “selling point” of Airdna but also of practical impor-
tance to prospective rental hosts, governments anxious to
understand the Airbnb phenomenon, and investors.3 An
increasing number of articles and studies rely on Airdna’s
measures of average daily rate (ADR), occupancy, and rev-
enue per available room (RevPAR).
This article examines whether Airdna’s performance
metrics are comparable with the norms of the traditional
lodging sector by examining the performance of Airbnb
rentals in Virginia Beach, Virginia. We compare the Airdna
metrics with those of STR and inquire whether Airdna
overstates the performance of the short-term lodging sec-
tor. If, as we argue, Airdna’s underlying definitions of
occupancy and RevPAR are biased upward, then Airbnb’s
impact is overstated—a conclusion of significant interest
to hoteliers, renters, investors, tax collectors, and policy-
makers alike.
The remainder of the article is structured as follows. In
the following section, we briefly review the extant literature
on the use of Airdna data to examine the performance of
Airbnb. We then discuss the potential flaws in the Airdna
methodology, with specific attention to the difference
between listing nights and room nights. We illustrate the
potential biases using Airdna data for the City of Virginia
Beach, Virginia. We then discuss Airdna’s response to our
queries and findings. The concluding section places our
findings in context and offers suggestions on how to correct
for these biases.
The Use of Airdna Data by
Researchers and Decision Makers
A growing number of popular press articles, consulting
studies, and peer-reviewed manuscripts rely on Airdna data
to estimate Airbnb activities. In some cases, Airdna data are
used to examine host characteristics, for example, whether
Airbnb hosts are comparable with those of free alternatives
such as “couch surfing” (Jung et al., 2016). Airdna data are
used to examine whether obtaining and maintaining repu-
table ratings is important for “superhosts” (Gunter, 2018).
In these cases, Airdna data may prove useful and reasonably
reliable as these studies rely more on characteristics of list-
ings rather than the performance of listings.
HVS Consulting and Valuation (2015) employed Airdna
data to estimate the impact of Airbnb on the New York City
lodging market and found that Airbnb poses a significant
threat to hoteliers’ revenues. CBRE Hotels’ Americas
Research (CBRE) found that Airbnb is not only growing
rapidly in major metropolitan areas but also that, in some
localities, Airbnb ADR exceeds hotel ADR (Lane &
Woodworth, 2016). CBRE also noted that in almost every
major market, the share of hosts with multiple units mark-
edly increased in 2016 and that revenue from hosts with
multiple units is growing faster than single unit hosts
(CBRE Hotel Americas Research, 2017). These, and other
consulting studies, rely extensively on Airdna data.
Airdna data have been used to examine the development
of Airbnb in the United States and other countries, to
include Austria (Gunter & Önder, 2017), Canada (Sovani
& Jayawardena, 2017), the Netherlands (Boswijk, 2017),
and South Africa (Visser, Erasmus, & Miller, 2017). Gibbs,
Guttentag, Gretzel, Yao, and Morton (2017), for example,
use Airdna listing data and STR data to argue that while
hotels employ dynamic pricing strategies to maximize rev-
enues, many Airbnb hosts in Canada do not employ a simi-
lar strategy.
Not surprisingly, listings for entire homes and private
rooms, listings with pictures, and host responsiveness com-
mand higher prices on Airbnb (Dogru & Pekin, 2017; Gunter
& Önder, 2017). Airdna occupancy data have been used as
the basis for an estimate that Asian and Hispanic hosts charge
8% to 10% lower prices in San Francisco (controlling for
rental unit characteristics, as well as additional information
on neighborhood property values and demographics; Kakar,
Voelz, Wu, & Franco, 2017). For Boston, Airdna data suggest
that Airbnb listings earn a significant price premium relative
to the long-term rental market (Horn & Merante, 2017).
A number of new studies directly compare Airdna and
STR performance data such as supply, ADR, and RevPAR
for Airbnb and hotels in Boston (Mody, Suess, & Dogru,
2017). For Los Angeles, Airbnb hosts’ ADR is higher than
the traditional lodging sector; however, occupancy and
RevPAR are lower than in the traditional sector (Dell,
Doby, Tillipman, & Zhuplev, 2017). The same authors
found that ADR and RevPAR are lower for Barcelona.
These studies employ Airdna’s performance metrics with-
out any adjustment.
Thus, Airdna data are being widely used for a variety of
purposes by many different parties. To the extent these data
do not impart the information their users believe to be the
case, the usefulness of studies and decisions based on these
data can be called into question.
Agarwal et al. 195
Are the Airdna Performance Measures
A central question relating to Airdna’s data is whether its
performance metrics are comparable with STR’s data,
which have constituted the industry standard for many
years. To understand the difference between the two sets of
data, we draw on the example of Virginia Beach. With
almost 450,000 residents, Virginia Beach is the most popu-
lous city in Virginia. The city is a well-known resort loca-
tion that annually attracts hundreds of thousands of visitors
to its beaches and other attractions. These visitors anchor
the city’s large conventional hotel and tourism sector. The
U.S. Travel Association estimated in 2016 that tourism was
responsible for US$1.494 billion of economic activity in
Virginia Beach (U.S. Travel Association, 2017). The perfor-
mance of the traditional lodging sector is of understandable
importance to hoteliers, tax collectors, and others associ-
ated with the tourist industry in Virginia Beach.
We purchased data from Airdna for the Virginia Beach
for the period of October 2014 through August 2017. The
data contained individual listings aggregated by month as
well as summary statistical information on various mea-
sures of performance by month. In January 2017, for
example, the data contained information on 421 listings
(the majority were residential homes) for Virginia Beach.
The individual listings varied from single private rooms to
significantly larger properties with as many as 10 bed-
rooms. The Airdna data contained a variety of market per-
formance measures that, at first glance, correspond to
published STR data.
Monthly Airdna data constitute a “list of listings.” Airbnb
classifies properties currently being offered for rent to
guests as available listings.4 Properties that are available
and have at least one reserved night in a month are consid-
ered booked listings. Unlike STR data that aggregate the
actual supply and bookings data of respondents, Airdna data
are estimates generated by a proprietary algorithm.
In the case of Virginia Beach, we agreed with Airdna’s
count of total available listings and total booked listings
when we examined listings of entire homes, private rooms,
and shared rooms.5 However, the first problem we noted
was that all properties that were available were not included
in the calculation of certain Airdna performance metrics.
This is a centerpiece of our concern. Simply being offered
for rent is not sufficient for a host’s property to be consid-
ered “available” by Airdna when it calculates occupancy
rates, ADRs, and RevPARs. Airdna includes only “booked
listings” in the calculation of these variables. This is equiva-
lent to a hotelier claiming that a room not rented during a
month should be excluded from the calculation of occu-
pancy rates and other measures.
In general, we found that the number of monthly avail-
able listings for one-bedroom homes exceeded the number
of actual booked listings. Thus, in August 2016, Virginia
Beach had 72 available listings for one-bedroom homes,
and of those, 64 were booked listings. This is a not a large
difference. During the offseason, however, the numbers
were markedly different. In February 2017, Virginia Beach
had 60 available listings for one-bedroom homes, but only
28 booked listings. On average, during the sample period in
question, there were 52 available listings for one-bedroom
homes compared to 37 booked listings each month. The
exclusion of properties without reservations clearly under-
states the available supply of Airbnb listings.6
Depending on the characteristics of the market in question,
the actual supply of Airbnb listings may be significantly higher.
In Virginia Beach, for example, the actual supply of Airbnb
listings was approximately 50% higher in some months when
all listings were included in the measure of supply. This is not
a trivial difference and constitutes significant bias.
A second problem emerged because of the way Airdna
populates the denominators of its major activity measures.
It employs total available nights as the denominator.
However, its total available nights are not, in many cases,
equal to the total number of available room nights, a subtle
but very important distinction. For listings that consist of
one bedroom (including shared rooms, private rooms, stu-
dios, and one-room entire houses), there is no distinction
between total available nights and total available room
rights. However, when the listing contains more than one
bedroom, then they are not identical. One listing may con-
tain multiple rooms.
With respect to Virginia Beach, we determined that the
ratio of booked room nights to booked nights was 2.11 for
the 35-month sample period. This average, however, con-
cealed a rapid increase in the ratio of actual room nights to
booked nights in late 2016 and into 2017, reaching a high of
2.88 rooms per listing in January 2017. The number of
available listings of properties with four or more bedrooms
increased significantly between 2015 and 2017.
We believe this reflects the increasing corporatization of
Airbnb properties. Anecdotal evidence from Virginia Beach
points to increasingly larger properties that are investor-
owned with the specific intent of using these properties for
short-term vacation rentals.7 The evolution of the ratio
strongly suggests that the trend in Virginia Beach is away
from single-bedroom properties. Given that listings of mul-
tiroom properties are proliferating on Airbnb, this consti-
tutes a bias that understates the number of available rooms
and, thus, Airbnb supply. This phenomenon is not limited to
Virginia Beach because investor-owned properties appear
to be rising as share of Airbnb properties nationwide.
Airdna data for Virginia Beach reveal that approximately
87% of Airbnb-derived revenues during our sample period
originated from apartments, condominiums, and full houses
rather than from properties that sought to rent out a single room
as part of a larger property. Of the revenues derived from
196 Cornell Hospitality Quarterly 60(3)
“whole property” rentals, approximately 68% originated from
properties with more than two bedrooms. We argue that this is
evidence of the increasing proliferation of multiroom proper-
ties and underscores the need to use available room nights as
opposed to Airdna’s current practice of available listing nights.
It could be, however, that some hotels have suites with
multiple rooms, and this would skew STR data in a similar
fashion. Our discussions with hoteliers revealed that hotel
suites generally are considered as one room, per industry
standard. We do not have information on the national distri-
bution of two-or-more bedroom suites to one-bedroom
suites or standard rooms, but we surveyed hotel owners in
the Hampton Roads region (officially the Virginia Beach–
Norfolk–Newport News Metropolitan Statistical Area) to
determine their numbers of multiroom suites. Our survey
included 69 hotels representing 7,782 rooms, or approxi-
mately 21% of all hotel rooms in the region. We found that
two-bedroom suites accounted for only 4.2% of all hotel
rooms among these surveyed hotels. There were zero two-
plus bedroom suites among the surveyed hotels.8
In sum, we conclude that Airdna’s performance mea-
sures are biased upward for two reasons. First, Airdna’s
metrics utilize booked listings. If a property is available
for rent during a given month, yet does not have any actual
reservations, it is excluded from Airdna’s measures of per-
formance. This methodology biases occupancy and
RevPAR upward. Second, even if the property in question
is available and rented for one (or more days), it is only
one listing by Airdna, regardless of the number of avail-
able bedrooms in the property in question. Our conclusion
is straightforward: Airdna’s estimates of ADR, occupancy,
and RevPAR for Airbnb are exaggerated. The actual num-
bers (i.e., the numbers based on standard STR definitions)
are lower.
Hence, Airbnb and Airbnb properties appear to be more
effective than they are in practice. The equivalent practice
for the traditional lodging sector would be to exclude rooms
that did not rent during a month from the calculation of
occupancy rates and RevPARs. This would have a signifi-
cant (and misleading) impact on these performance metrics,
especially in tourist-centric areas that typically operate
much slower during off seasons. These methodological
flaws mean that it is difficult to make meaningful compari-
sons between Airdna and STR data. The two firms are not
measuring the same things—it is a classic “apples and
oranges” circumstance.
This makes it much more difficult to assess the relative
importance and performance of Airbnb properties versus those
of hotels and motels. Comparable metrics are required to do so.
The publicly available Airdna market performance data are not
comparable with STR data. For those interested in precision
and accuracy, their only current recourse is to use Airdna indi-
vidual listing data to calculate performance metrics that are
comparable with the industry standard. Otherwise, they may
rely on performance data that are biased upward and draw spu-
rious conclusions as to the health of the Airbnb marketplace.
Discussions With Airdna
There is nothing dishonest about the way Airdna chooses to
compute occupancy rates, ADRs, and RevPARs. Nevertheless,
Airdna has chosen not to employ commonly accepted defini-
tions of these performance metrics, and thus, its data will be
misleading when unknowing individuals attempt to compare
the Airdna performance information with that of STR.
Given our concerns regarding the methodological foun-
dations of the Airdna performance metrics, we initiated a
conversation with Airdna regarding the calculation of the
performance metrics. The first response from Airdna was,
Can you elaborate on your concerns for these metrics? We have
been covering data on these locations for a significant period of
time (30 months and counting in the US) and provided data to
almost 20,000 customers from individual hosts/investors
through to multinational organizations and over fifty globally
recognized research institutions such as Harvard, Oxford,
Princeton and National University of Singapore so our data has
been thoroughly carved up and crunched. To date, we have not
had any wide-ranging concerns on our methodology from any
of these sources so we would be keen to hear your feedback in
greater detail.9
On receipt of this communication, we asked two specific
questions. Our first question highlighted the first potential
source of bias, that is, the difference between total booked
nights and total booked room nights. We wrote,
. . . when Airdna reports ADR for the entire place, you are
using as denominator your measure of total booked nights
(regardless of the size of the property). The denominator is not
the total number of booked room nights. This is a meaningless
distinction where one bedroom homes are concerned because
nights equal rooms, but makes a significant difference for the
entire place when one recognizes that many Airbnb hosts rent
homes with multiple bedrooms.
We received the following email response:
We considered different ways to report ADR and we found
using “Entire Place” to be the most helpful not looking at the
ADR per room within an “entire home listing.” If you wish to
calculate this ADR you have the possibility doing so by using
the property file.
Our second question was,
In calculating available nights or available room nights, Airdna
excludes from consideration any property that might have been
available for rent, but had no reservations in the month in
Agarwal et al. 197
question. This exclusion means Airdna is underestimating
available nights or available room nights resulting in
overestimation of occupancy rates and RevPAR.
Airdna’s response was,
About our occupancy rate calculation, when we designed our
reports we had a lengthy internal discussion about this. Our
CEO had many Airbnb properties and knows that if someone is
actively managing their property they should be able to have a
reservation every month. We decided on the annual occupancy
rate calculation not to take into consideration months with no
booking as throwing lots of un-managed, zero occupancy,
properties into the calculation doesn’t give a good picture of
In effect, Airdna’s collective responses were, “Lots of
people are using our data. They aren’t complaining.
What’s the problem?” Nevertheless, we suspect that firms,
governments, and researchers would view Airdna data dif-
ferently if they were aware of the measurement differ-
ences we have highlighted here. We appreciate Airdna
providing us with insight into its methodological decisions
underlying its performance metrics because the firm did
not have to do so. Even so, its practices have made its
performance metrics incompatible with industry stan-
dards. Airdna metrics are biased upward and provide a
rosier picture of the health (and potential profitability) of
Airbnb properties than would otherwise be provided if
Airdna chose to adapt their methodology to widely
accepted practices.
Final Observations
The Bureau of Economic Analysis (BEA) of the U.S.
Department of Commerce defines and computes Gross
Domestic Product (GDP). Using the same broad defini-
tion—the total value of the goods and services produced
in each jurisdiction during a given period, usually a
year—the BEA also computes GDP at the state and met-
ropolitan area level. Suppose now that another organiza-
tion began to produce estimates of GDP for specific cities,
but in doing so decided to alter the BEA definition by
adding the value of leisure time. Would we consider these
new numbers to be comparable with the widely accepted
BEA estimates? We are confident the answer is no.
The same reasoning applies to Airdna’s performance
measures. Life would be easier for the hotel and motel
industry, cities and counties, elected officials, Airbnb
hosts, and researchers if Airdna adhered to standard,
widely accepted definitions. We recommend that Airdna
change its methodology in two specific ways. When
determining market supply, Airdna should employ avail-
able listings rather than booked listings. If a property is
available for rent, then it should not be excluded from
analysis merely because it failed to secure a rental during
a given month. Given that Airbnb supply was underesti-
mated by approximately 50% in Virginia Beach in some
months, this is an important adjustment to make to ensure
that consumers of Airdna data have an accurate measure
of Airbnb supply.
Second, Airdna should use available room nights rather
than available nights when calculating ADR and RevPAR.
As the number of multiroom properties increases, the bias
of calculating based on listings will undoubtedly increase.
While this adjustment lowers the estimated performance of
Airbnb rentals as we found that there was an average of two
rooms per listing in Virginia Beach, it is necessary if one
desires to compare and contrast the performance of the
short-term rental sector with the traditional lodging sector.
In the meantime, we recommend that individuals inter-
ested in using the Airbnb data use the individual listing data
to make the adjustments that we have discussed in this arti-
cle. In the end, we should neither formulate public policies
nor make private investment decisions on the basis of apples
and oranges comparisons.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with
respect to the research, authorship, or publication of this article.
The author(s) received no financial support for the research,
authorship, or publication of this article.
1. Airbnb launched an official Application Programming
Interface (API) in October 2017. The API is not open as
interested parties must apply for access, and it appears to be
focused on property owners.
2. Note that Airdna currently captures and sells data for Airbnb
but does not offer similar data for other vacation rental web-
sites such as Flipkey, Homeaway, or Vacation Rentals By
Owner (VRBO) .
3. According to Airdna, the market summary reports “provide a
high level overview of Airbnb rentals in major markets around
the world. With revenue per available room (RevPAR), average
daily rate (ADR), Occupancy, and monthly supply & demand
trends.” Airdna claims these reports “bring hotel-style perfor-
mance metrics to the vacation rental industry” (Airdna, 2017).
4. Each individual observation contains a listing status identifier
that categorizes the listing status as either “true” or “false.” A
listing property’s status is false if all the days in a month are
blocked days (i.e., unavailable to guests) or if the Airbnb cal-
endar for the property in question was not visible on Airbnb for
that month. A listing’s property status is true if there is at least
one day a month that is not blocked, or if the Airbnb calendar
for the property is visible and the owner is accepting reserva-
tions. A status of true implies that a listing is currently being
198 Cornell Hospitality Quarterly 60(3)
offered for a guest to rent or that it has been rented by a guest in
each month.
5. Subsequently, we chose to exclude shared rooms because
the total number of such listings was very small; the number
never exceeded nine, and the single room share of total avail-
able listings never exceeded 2.44%.
6. Smith Travel Research, “A Guide to Our Terminology,” www. For example, Smith
Travel defined occupancy (OCC) as the “percentage of available
rooms sold during a specified time period. Occupancy is calcu-
lated by dividing the number of rooms sold by rooms available.”
7. In our discussions with representatives of the Virginia Beach
Commissioner of Revenue and that city’s Convention and
Visitors Bureau, it became apparent that several new devel-
opments of apartments and condominiums were investor-
owned and targeted for short-term rentals.
8. Survey questions and aggregated responses are available on
9. Email received from Adam Alexander of Airdna on May 10,
10. Email received from My Larson of Airdna on June 16, 2017.
Airbnb. (2017). About us—Airbnb. Retrieved from https://press.
Airdna. (2017). Datafeed / API services. Retrieved from https://
Boswijk, A. (2017). Transforming business value through digi-
talized networks: A case study on the value drivers of Airbnb.
Journal of Creating Value, 3, 104-114.
CBRE Hotel Americas Research. (2017). Hosts with multiple
units—A key driver of Airbnb growth. Retrieved from https://
Dell, J., Doby, D., Tillipman, J., & Zhuplev, A. (2017). The
impacts of the peer-to-peer platform on the traditional lodg-
ing industry: Emerging trends and implications for greater
Los Angeles (U.S.A) and Barcelona (Spain). The Journal of
Applied Business and Economics, 19, 130-158.
Dogru, T., & Pekin, O. (2017). What do guests value most in
Airbnb accommodations? An application of the hedonic pric-
ing approach. Boston Hospitality Review, 5(2), 1-3.
Edelman, B. G., & Geradin, D. (2016). Efficiencies and regulatory
shortcuts: How should we regulate companies like Airbnb and
Uber? Stanford Technology Law Review, 19, 293-328.
Fang, B., Ye, Q., & Law, R. (2016). Effect of sharing economy on
tourism industry employment. Annals of Tourism Research,
57, 264-267.
Gibbs, C., Guttentag, D., Gretzel, U., Yao, L., & Morton,
J. (2017). Use of dynamic pricing strategies by Airbnb
hosts. International Journal of Contemporary Hospitality
Management. Retrieved from http://www.emeraldinsight.
Gunter, U. (2018). What makes an Airbnb host a superhost?
Empirical evidence from San Francisco and the Bay Area.
Tourism Management, 66, 26-37.
Gunter, U., & Önder, I. (2017). Determinants of Airbnb demand in
Vienna and their implications for the traditional accommoda-
tion industry. Tourism Economics, 24(3), 270-293.
Guttentag, D. A. (2015). Airbnb: Disruptive innovation and the
rise of an informal tourism accommodation sector. Current
Issues in Tourism, 18, 1192-1217.
Guttentag, D. A., & Smith, S. L. J. (2017). Assessing Airbnb as
a disruptive innovation relative to hotels: Substitution and
comparative performance expectations. International Journal
of Hospitality Management, 64, 1-10.
Horn, K., & Merante, M. (2017). Is home sharing driving up
rents? Evidence from Airbnb in Boston. Journal of Housing
Economics, 38(Suppl. C), 14-24.
HVS. (2015). Impact analysis report: Airbnb and impacts on the
New York City lodging market and economy. Retrieved from
Jordan, G. (2015, August 26). Airbnb not just a worry, but “tec-
tonic shift.” Hotel News Now. Retrieved from http://www.
Jung, J., Yoon, S., Kim, S., Park, S., Lee, K.-P., & Lee, U. (2016).
Social or financial goals? Comparative analysis of user
behaviors in couchsurfing and Airbnb. In Proceedings of the
2016 CHI Conference Extended Abstracts on Human Factors
in Computing Systems (pp. 2857-2863). Retrieved from doi.
Kakar, V., Voelz, J., Wu, J., & Franco, J. (2017). The visible
host: Does race guide Airbnb rental rates in San Francisco?
Journal of Housing Economics, In Press, Retrieved from doi.
Kirkham, C. (2017, July 24). Expedia, priceline home in on
Airbnb’s turf. Retrieved from
Koopman, C., Mitchell, M. D., & Thierer, A. D. (2015). The shar-
ing economy and consumer protection regulation: The case
for policy change (SSRN Scholarly Paper No. ID 2535345).
Rochester, NY: Social Science Research Network.
Lane, J., & Woodworth, R. M. (2016, January). The sharing
economy checks in: An analysis of Airbnb in the United
States. CBRE. Retrieved from
Mody, M., Suess, C., & Dogru, T. (2017). Comparing apples
and oranges? Examining the impacts of Airbnb on hotel
performance in Boston. Boston Hospitality Review, 5(2),
Neeser, D., Peitz, M., & Stuhler, J. (2015). Does Airbnb hurt hotel
business: Evidence from the Nordic counties. University
Carlos II de Madrid. Retrieved from
Quattrone, G., Proserpio, D., Quercia, D., Capra, L., & Musolesi, M.
(2016). Who benefits from the “sharing” economy of Airbnb?
Arxiv. Retrieved from
Sovani, A., & Jayawardena, C. (2017). How should Canadian
tourism embrace the disruption caused by the sharing
economy? Worldwide Hospitality and Tourism Themes, 9,
U.S. Travel Association. (2017, September). 2016 economic
impact of domestic travel to Virginia and localities. Virginia
Tourism Corporation, Richmond, Virginia Retrieved from
Agarwal et al. 199
Visser, G., Erasmus, I., & Miller, M. (2017). Airbnb: The emer-
gence of a new accommodation type in Cape Town, South
Africa. Tourism Review International, 21, 151-168.
Zervas, G., Proserpio, D., & Byers, J. W. (2017). The rise of the
sharing economy: Estimating the impact of Airbnb on the hotel
industry. Journal of Marketing Research, 54(5), 687-705.
Author Biographies
Vinod Agarwal is a Professor of Economics in the Deparmtent of
Economics in the Strome College of Business at Old Dominion
University. He also serves as the Director of the Economic
Forecasting Project and as deputy Director of the Dragas Center
for Economic Analysis and Policy.
James V. Koch is an Emeritus Professor of Economics at Old
Dominion University.
Robert M. McNab is a Professor of Economics in the
Department of Economics in the Strome College of Business at
Old Dominion University. He also serves as the Director of the
Dragas Center for Economic Analysis and Policy.
... Finally, information from AirDNA (2020) was used to estimate the volume of housing available for tourist use in Ibiza and Formentera to supplement the data on tourist accommodation published by IBE-STAT (2020). AirDNA collects information from Airbnb and has been used in previous studies (e.g., Gunter et al., 2020;Ioannides, R€ oslmaier, & van der Zee, 2019), although this data should be used with caution (Agarwal, Koch, & McNab, 2019). ...
... For example, the following numbers of tourist-use dwellings were all found: 17,153 locations on Airbnb in 2016 (Groizard & Nilsson, 2017); 30,018 locations on Airbnb in , 29,894, in 2018(Terraferida, 2017), 42,670 in 2017(Falc on & Palma, 2018; and 27,189 places on Airbnb in 2020 (AirDNA, 2020). Based on this finding and those of previous research (Agarwal et al., 2019), the data on available tourist-use housing should be interpreted with great caution. ...
Full-text available
In the last five years, home rentals for tourists have generated significant media and social concern. Since the economic crisis of 2007–2008, online platforms have emerged that have boosted the collaborative economy and provided security and trust. There are two positions related to the impacts of peer-to-peer and intermediation platforms: some consider them to reflect the destruction of tourist destinations at the hands of international corporations, while others allege that they have helped improve the incomes of many families and meet demand. These positions raise doubts, and based on the descriptive analysis of secondary public data from Spain, this study makes a first approximation of the actual situation. The results indicate that the platforms have given online visibility to business that already existed, while tourist rentals cater to specific market niches and adjust hotel supply to tourist demand. The study concludes that tourist rentals have been confused with the real problem: large and rapid increases in demand that are difficult to manage, aggravating mismatches in the rental market that are the result of multiple factors.
... A significant number of scholars publishing in peer review journals (Agarwal, Koch, & McNab, 2019) and consultancies (CBRE Hotel Americas Research, 2017;Kelley & Asad, 2015) appear to rely on AirDNA for estimating Airbnb activities. In many cases, data from AirDNA has been used to survey host features in comparison to couch surfing (Jung et al., 2016). ...
... Following existing research that used AirDNA for data analysis, such as Agarwal et al. (2019); CBRE Hotel Americas Research (2017); Gunter (2018) and Kakar et al. (2017). In examining Airbnb performance, we used AirDNA dataset between May 2017 to April 2020 from the following 17 cities in the Cape Town neighbourhood, Bloubergstrand, Cape Town City Centre, Sea Point, Greenpoint, Hout Bay, Camps Bay, Somerset West, Gardens and Constantia. ...
Full-text available
Some economies are shrinking, while others are experiencing downtime due to the spread of the coronavirus disease pandemic. Social integration and rural development are weakened in municipalities in South Africa as a result of government mediated approach for reducing the curve of the Coronavirus disease, such as a lockdown. This paper assesses the effect of COVID-19 lockdown on Airbnb performance between May 2017 and April 2020 in 17 cities in Cape Town neighbourhood with a focus on the COVID-19 era January 1-April 30, 2020. Our finding demonstrates that COVID-19 did not have a significant effect on Airbnb performance in Cape Town. However, policies for flattering the COVID-19 curve, such as lockdown which brought about travel restrictions, depleted international and local guests significantly and reduced the revenue of Airbnb by over 200 million Rands in a month March to April 2020 in 17 cities in the Cape Town neighbourhood.
... Airdna is widely considered to be a leading purveyor of Airbnb listing data and provides a valuable service to those without direct access to Airbnb listing data. Questions, however, remain about how Airdna calculates its measures of Average Daily Rate (ADR), Occupancy, and Revenue per Available Room (RevPAR), whether these measures conform to industry standard, and the extent of bias (if any) (Agarwal et al., 2019). 2 In this paper, we replicate and extend the literature with regards to Airdna's performance measures. ...
Full-text available
Because individual listing data for Airbnb typically are not publicly available, private companies have emerged to estimate the performance of Airbnb listings. The implicit assumption of a growing number of academics, policymakers, and consultants is that Airdna’s performance measures are directly comparable with those of STR. We argue that Airdna’s measures of Occupancy, Average Daily Rate (ADR), and Revenue per Available Room (RevPAR) do not conform to industry standards and exhibit significant bias. We expand available evidence by explicitly quantifying the sources and magnitude of the biases for Airdna’s performance measures for Airbnb listings. Using Airdna’s individual listing data for Virginia between the first quarter of 2015 and the 4th quarter of 2019, we find, on average, Airdna’s performance measures for Occupancy, ADR, and RevPAR were biased upward by 60 percent, 78 percent, and 179 percent, respectively.
... Another concern is the quality of the AirDNA data. Agarwal, Koch, and McNab (2018) investigated data from Virginia Beach, Virginia, and they argued that AirDNA does not utilize the Smith Travel Research (STR) standard to collect their data, which might induce bias in our estimates. ...
The increasing number of natural hazards in coastal tourism destinations has negatively affected their local lodging industries. The recent boom in shared accommodation in coastal destinations is also under threat due to increasing sea levels and extreme weather events. Thus, estimating the economic impact of natural hazards on shared accommodation is a critical prerequisite for effective tourism destination crisis management. This study aimed to estimate the economic impact of natural hazards on shared accommodation. To achieve this purpose, HAZards U.S. MultiHazard (HAZUS-MH) hurricane/flood models were employed in a case study of Hurricane Irma and 822 Airbnb properties in Collier County, Florida, for 2017. The estimated direct combined losses from wind gusts and storm surge flooding were $22,683,054, and the indirect losses to rental income were as high as $19,120 per day. This estimation method can help individual owners and local government managers predict problems related to natural hazards and more effectively prepare for future hazardous events in coastal tourism destinations.
Zusammenfassung Die Interaktion von Tourist:innen und Einheimischen und damit die Akzeptanz von Tourismus ist verstärkt in den Blick von Tourismusforschung und Destinationspraxis geraten. Studien zeigen, dass fehlende Tourismusakzeptanz ein punktuelles Problem ist und dass die Übernachtungsintensität einen Einfluss auf die Reaktion von Einheimischen hat. Bislang fehlen jedoch flächendeckende Daten zur Übernachtungsintensität und -konzentration der Gemeinden in Deutschland. In dieser Studie werden erstmals die statistisch erfassten Übernachtungsangebote und zusätzlich die Angebote in den Buchungssystemen von Airbnb und Fewo-direkt mit dem Ziel zusammengeführt, eine gemeindescharfe Abbildung der Übernachtungskonzentration und -intensität zu erarbeiten. Die Ergebnisse zeigen zum einen, dass 81 % der Gemeinden in Deutschland übernachtungstouristische Angebote vorhalten. In diesen Gemeinden leben 98 % der Einwohnerinnen und Einwohner. Tourismus geht also fast alle an. Die Ergebnisse zeigen außerdem, dass Privatvermietungen nicht zu einer Entzerrung der räumlichen Konzentration führen, das Konzentrationsmaß ist bei Privatvermietungen höher als bei statistisch meldepflichtigen Betrieben.
In this paper we aim to classify digital data sources for the measurement of tourist mobility, to establish a set of assessment indicators, and to compare two Big Data sources to gain empirical insights into how we can measure tourism with Big Data. For three holiday destinations in Germany, passive mobile data and passive global positioning systems (GPS) data are compared with reference data from the destinations for twelve weeks in the summer of 2019. Results show that mobile network data are on a plausible level compared to the local reference data and are able to predict the temporal pattern to a very high degree. GPS app-based data also perform well, but are less plausible and precise than mobile network data.
This study seeks to quantify the value of coastal amenities to the blue tourism economy. It assesses the value of recreational and aesthetic coastal amenities in an area where tourism-related development is a key economic activity, specifically, the West Flemish vacation rental sector. Using spatial panel data comprising a cross section of 821 active Airbnb listings acquired from AirDNA, covering the time period from March 2019 to February 2020, the study models the effects of sea view and coastal proximity on Airbnb revenue alongside other structural and locational attributes. Rental units with a view of and close proximity to the sea yielded significantly higher revenues.
This conceptual paper puts forth sources of secondary and primary data that can assist destinations in developed countries track and ultimately improve their communities' tourism performance and resilience across four pillars of tourism sustainability: visitor economy, resident support, workforce satisfaction, and environmental health. Such data sources can be used by community leaders to formulate key performance indicators (KPIs) critical in building resilience. The impetus behind these proposed solutions is based on the assumption that the high cost and complexity of collecting and analyzing data on an ongoing basis have plagued the broad adoption of communities engaging in sustainable initiatives. We contend that given access to suitable data, destinations can be better managed as ecosystems in line with sustainable community objectives.
Conference Paper
Full-text available
Purpose – This paper aims to identify major supply data sources for short-term rental market research and to provide their advantages and limitations. Methodology – In the paper a grounded approach was used based on a literature review. This review comprised two steps with the first being the query in major databases that was supplemented by academic search engine that resulted in 170 articles. The second step was to investigate the papers’ methodological sections to identify characteristics and limitations of all data sources. Findings – This study identifies three major data sources for the short-term rental market: web scraping with the use of self-made bots, Inside Airbnb and Airdna. A majority (e.g. 74% of papers using Airdna as a source) did not mention any limitations and provide no discussion about the data source, while the remainder gave only superfluous information about possible limitations of its use. Their characteristics and limitations are extensively discussed using a proposed framework that consists of three levels: intermediary, web scraping, and source-specific. Contribution – Very limited number of studies have focused on the short-term rental data sources and this is the first one that discusses advantages and limitation of their use. This paper may be of help to academics or professionals in identifying the right source of data to suit their technical knowledge, financial and technical resources and research areas.
Extant research provides ample evidence that Airbnb has an adverse impact on hotel revenues and that a majority of Airbnb hosts offer multiple listings on the platform. Arguably, multi-unit host listings are the primary driving force for the associated decreases in hotel revenues. To test this proposition, this study examines the extent to which single- and multi-unit host Airbnb host listings have differing effects on hotel revenues. The results show that the adverse impact of Airbnb supply on hotel revenues was mainly driven by decreases in hotel prices rather than decreases in hotel demand. Contrary to expectations, single-unit host listings exerted greater downward pricing pressure on hotels than multi-unit host listings. These effects were consistently observed across hotel segments, with economy-scale hotels experiencing substitution effects in addition to pricing pressure. The findings have significant implications for researchers and stakeholders impacted by the sharing economy in the lodging industry.
Full-text available
Purpose The purpose of this paper is to provide a comprehensive analysis of dynamic pricing by Airbnb hosts. Design/methodology/approach This study uses attribute and sales information from 39,837 Airbnb listings and hotel data from 1,025 hotels across five markets to test different hypotheses which explore the extent to which Airbnb hosts use dynamic pricing and how their pricing strategies compare to those of hotels. Findings Airbnb is a unique and complex platform in terms of dynamic pricing where hosts make limited use of dynamic pricing strategies, especially as compared to hotels. Notwithstanding their limited use, hosts who own listings in high-demand leisure markets, manage entire places, manage more listings and have more experience vary prices the most. Practical implications This study identified a great need for Airbnb to encourage dynamic pricing among its hosts, but also warned of the potential perils of dynamic pricing in the sharing economy context. The findings also demonstrated challenges for hotel managers interested in actionable information related to Airbnb as a competitor. Originality/value This is the first Airbnb study to use a comprehensive set of data over a continuous period in multiple markets to look at a number of listing and host factors and determine their relation with dynamic pricing strategies.
Full-text available
The sharing economy phenomenon, and the economic, social, and technological changes fueling its growth have challenged the hotel industry to rethink its experiential value proposition to the customer. But, is Airbnb taking a share of the existing hotel industry or is it just increasing the overall accommodations industry?
Using data on Airbnb listings from San Francisco and the Bay Area, the present study investigates the relative importance of the four criteria that need to be fulfilled to obtain the Airbnb superhost status. In order to quantify the marginal contributions of the four criteria, different index models of binary response (logit, probit, and IV probit, which allows for the endogeneity of Airbnb demand) are applied. The results, which are consistent across models, show that in San Francisco and the Bay Area obtaining (and maintaining) excellent ratings is, by far, the most important criterion, followed by reliable cancellation behavior of the host, host responsiveness, and sufficient Airbnb demand. Moreover, commercial Airbnb providers are more likely to obtain the superhost status.
The growth of the sharing economy has received increasing attention from economists. Some researchers have examined how these new business models shape market mechanisms and, in the case of home sharing, economists have examined how the sharing economy affects the hotel industry. There is currently limited evidence on whether home sharing affects the housing market, despite the obvious overlap between these two markets. As a result, policy makers grappling with the effects of the rapid growth of home sharing have inadequate information on which to make reasoned policy decisions. In this paper, we add to the small but growing body of knowledge on how the sharing economy is shaping the housing market by focusing on the short-term effects of the growth of Airbnb in Boston neighborhoods on the rental market, relying on individual rental listings. We examine whether the increasing presence of Airbnb raises asking rents and whether the change in rents may be driven by a decline in the supply of housing offered for rent. We show that a one standard deviation increase in Airbnb listings is associated with an increase in asking rents of 0.4%.
Peer to Peer e-commerce is increasingly characterized by trends towards the personalization of buyers and sellers in the on-line marketplace. This personalization includes buyer reviews, personal pictures and profiles, and other biographical information intended to reduce buyers' perceived "purchase risk" or to facilitate trust in the sellers. However, this phenomenon is transforming what started as an essentially "anonymous" market to one susceptible to traditional market failures, including potential racial discrimination, in a manner similar to its brick and mortar counterparts. In this paper, we examine the effect of on-line host information (race, gender, sexual orientation, etc.) on the price of available rental listings in San Francisco on We find that on average, Asian and Hispanic hosts charge 8%-10% lower prices relative to their White counterparts on equivalent rental properties, after controlling for all renter-available information on rental unit characteristics, as well as additional information on neighborhood property values, area demographics, and occupancy rates. We do not find any differences in occupancy rates between minority and White hosts. This may suggest that minorities price lower because they are forward-looking, perhaps due to an expectation of discrimination in the online marketplace or have a preference to increase demand to either maintain their target occupancy level or to attract a larger pool of potential renters to choose from. Overall, our findings are consistent with but not conclusive of a market test of potential racial discrimination affecting Hispanic Airbnb hosts, manifested in an anticipation of disparate market demand for their rentals, and responded to by lower listing prices.
This study identifies key determinants of Airbnb demand and quantifies their marginal contributions in terms of demand elasticities. A comprehensive cross-sectional data set of all Viennese Airbnb listings active between July 2015 and June 2016 is examined. Estimation results, which are obtained by cluster-robust OLS, show that Airbnb demand in Vienna is price-inelastic. Significant positive drivers include listing size, number of photos, and responsiveness of the host. Significant negative drivers include listing price, distance from the city center, and response time of the host. Implications for the traditional accommodation industry are that, on the one hand, it should better communicate its sought-after advantages (e.g. lower average minimum duration of stay). On the other hand, it should increase its offer of bigger and better equipped hotel rooms since hosting more than two guests at a time is one of the major benefits of Airbnb.
The appearance and expansion of Airbnb as part of the sharing economy have generated a growing debate in tourism scholarship. Within sub-Saharan Africa, South Africa is currently the major focus of Airbnb operations. This article investigates the recent proliferation of Airbnb's accommodation in Cape Town, South Africa's premier international tourist destination. Responding to changing tourist preferences and dynamic change in readily accessible technologies, Airbnb has found a niche in the tourist accommodation market left vacant by more traditional and formal tourist accommodation types. Despite its explosive expansion, little is known about this phenomenon in South Africa. This investigation seeks to address the limited scholarship on this tourist accommodation segment in South Africa in the empirical setting of Cape Town. Issues under scrutiny are the spatial and temporal development of Airbnb in Cape Town and the characteristics of Airbnb providers and guests.
Purpose This paper aims to answer two questions: 1. What is the sharing economy? 2.How is the sharing economy affecting tourism in Canada? Design/methodology/approach The foundation of this paper was laid during a major industry event held in Ottawa in 2016 – the Ontario Tourism Summit, an annual industry conference organized by the Tourism Industry Association of Ontario (TIAO), attended by 650 industry participants. This article is based on presentations made at the summit. The article provides key information on Airbnb and the role of TIAO in the context of shared economy. Findings Companies such as Airbnb, Uber and Turo have made the concept of sharing economies an everyday concept. As sharing economy is considered as a phenomenon here to stay, Canadian tourism and hospitality industries should embrace the disruption caused by it and ensure that this is done for mutual benefit of all stake holders. Five key suggestions are made by authors in their conclusions. Practical implications As this article is mainly based on author’s viewpoints, prior to implementing their recommendations, further dialog with all relevant stake holders is needed. Originality/value This paper draws upon the author’s experience working with Canadian tourism companies and incorporates their thoughts for practical solutions.