Surprising Airbnb Adoption Slowdown in US/EU, and What It Means for Hotels and OTAs

Transcription

1 November, 7 5:7 A GT Internet, Lodging, Leisure Surprising Adoption Slowdown in US/EU, and What It eans for Hotels and OTAs O ur 3rd annual AlphaWise survey suggests adoption is slowing in the US/ Europe as awareness peaks and privacy/safety issues rise. Hotel cannibalization is up modestly but slower adoption is positive for hotels, so we raise RevPAR. OTAs are becoming stronger competitors against. organ Stanley does and seeks to do business with companies covered in organ Stanley Research. As a result, investors should be aware that the firm may have a conflict of interest that could affect the objectivity of organ Stanley Research. Investors should consider organ Stanley Research as only a single factor in making their investment decision. For analyst certification and other important disclosures, refer to the Disclosure Section, located at the end of this report. + = Analysts employed by non-u.s. affiliates are not registered with FINRA, may not be associated persons of the member and may not be subject to NASD/NYSE restrictions on communications with a subject company, public appearances and trading securities held by a research analyst account.

2 Contributors ORGAN STANLEY & CO. LLC ORGAN STANLEY & CO. INTERNATIONAL PLC+ Brian Nowak, CFA Jamie Rollo Equity Analyst Equity Analyst Brian.Nowak@morganstanley.com Jamie.Rollo@morganstanley.com ORGAN STANLEY & CO. LLC ORGAN STANLEY & CO. LLC ORGAN STANLEY & CO. INTERNATIONAL PLC+ Thomas Allen ark Savino Adrija Chakraborty Equity Analyst Equity Analyst Equity Analyst Thomas.Allen@morganstanley.com ark.savino@morganstanley.com Adrija.Chakraborty@morganstanley.com ORGAN STANLEY & CO. LLC ORGAN STANLEY & CO. LLC ORGAN STANLEY & CO. LLC Alaxandar Wang Chaodan Zheng atthew Cost Research Associate Research Associate Research Associate Alaxandar.Wang@morganstanley.com Chaodan.Zheng@morganstanley.com atthew.cost@morganstanley.com

3 Contents 6 Key Takeaways Key Debate #: How Quickly Is User Adoption Growing? 6 Key Debate #: How Big of a Competitive Threat Is to Hotels? 3 Key Debate #3: How Big a Competitive Threat Is to OTAs? 9 Appendix 33 odel and Impact to Hotel RevPAR 35 What Is? 37 ore on AlphaWise ORGAN STANLEY RESEARCH 3

4 Internet, Lodging, Leisure Surprising Adoption Slowdown in US/EU, and What It eans for Hotels and OTAs O ur 3rd annual AlphaWise survey suggests adoption is slowing in the US/ Europe as awareness peaks and privacy/safety issues rise. Hotel cannibalization is up modestly but slower adoption is positive for hotels, so we raise RevPAR. OTAs are becoming stronger competitors against. Internet North America Industry View Gaming & Lodging North America Industry View Leisure and Hotels Europe Industry View Attractive In-Line In-Line Adoption Curve in US and Europe Flattening Sooner Than We Expected: Our AlphaWise survey data indicates adoption has slowed, as the percent of travelers who used over the last months increased by only ~33bps to, a notable deceleration from the ~8bps increase in 6. Two Reasons Adoption Is Slowing... First, the benefits of growing awareness among online consumers in the US/Europe are slowing/topping out, as awareness among our survey respondents increased by only 8bps to in '7 (vs. a,bps increase in '6). We see awareness as a smaller driver going forward. Second privacy/safety are material and growing barriers to adoption, as the percent of non users citing these factors to not use it increased by ~7bps/4bps, and the absolute number of people concerned about these issues increased by %/ Y/Y. This, in our view, could speak to two potential truths: ) How /sharing lodging could be more niche than previously expected, and ) How the lobbyists/opponents of may be gaining traction....causing Us to Lower Our Forecast...This slowing adoption causes us to reduce our forward user/room night forecast for the US/Europe, now expecting ~% '6-' room night growth, vs 9% previously. We now model to grow to of lodging demand across US/Europe by ', vs. 9% previously....aking Us Incrementally ore Bullish on the Lodging Space... While surveyed hotel cannibalization inched up slightly to 5% from 49% (and expected to be 5 in '8), a slowing user adoption curve suggests is less of a threat to hotels. We estimate that had a 5bps impact to '7 RevPAR growth across US/Europe, down from our prior estimate of 9bps. We now forecast it will have another 5bps impact on '8 RevPAR growth, down from our prior 8bps impact. As such, we expect US RevPAR to be relatively stable at +. and +. growth in '8 and '9, respectively. organ Stanley does and seeks to do business with companies covered in organ Stanley Research. As a result, investors should be aware that the firm may have a conflict of interest that could affect the objectivity of organ Stanley Research. Investors should consider organ Stanley Research as only a single factor in making their investment decision. For analyst certification and other important disclosures, refer to the Disclosure Section, located at the end of this report. += Analysts employed by non-u.s. affiliates are not registered with FINRA, may not be associated persons of the member and may not be subject to NASD/NYSE restrictions on communications with a subject company, public appearances and trading securities held by a research analyst account.

5 We raise our Lodging stocks' EBITDA forecasts, and given implied stability, shift our price targets to be based off '9 EBITDA, which pushes our price targets up on average and implies average upside (including dividends). In addition, with less of a structural threat, our bull and bear values move higher, shifting the risk-reward skew from even to more positive....and the OTAs, Who Are (Arguably) Becoming Bigger Threats to. OTA usage is still rising as OTAs/ are the only two booking channels up Y/Y. This speaks to the resiliency/strength of the OTA model and consumer value proposition. If anything, we see the OTAs' strong use trends, paid search/traffic acquisition expertise and improving inventory offerings (more alternative accommodations) becoming a growing threat to, as the OTAs, in our view, are better positioned to act as the online travel one-stop shop.

6 Key Takeaways Exhibit AlphaWise Survey ethodology This note showcases our third annual AlphaWise survey around usage. We use this proprietary survey of Our AlphaWise survey shows of total travelers used in the last months...lower vs. expectation of in last year's survey, indicating respondents may be overestimating future use Use (Total Travelers) ~4, consumers in US and Europe in order to track adoption trends by country and the key drivers and hurdles 3 to growth. Over the past three years, each October- evenly distributed across the US, UK, France, and Germany. % 9% +788bp November we have surveyed ~4, adults aged 8+ The sample is representative of the online adult population +39bp % in terms of age, gender, income, and region in each country. We posed a series of questions around, leisure/ % corporate travel, travel accommodation behavior, and demographic data. Conclusions based on the total sample % have an estimated margin of error of +/-. at a 9% 5 6 Total Travelers 7A 7E 8E Source: AlphaWise, organ Stanley Research confidence level. Our findings are as follows. This trend holds both for leisure and business travelers, with actual adoption over the last months roughly 5bp below last year's forward expectation for both types of travelers (see Exhibit ) New Survey Data Speak to Slowing Adoption in US/ ). Leisure travel adoption increased from 9% in 6 to % in 7 Europe... (vs. expected adoption of ); and business adoption grew from in 6 to % in 7 (vs. expected adoption of ). Looking According to our survey, in the past year, adoption growth forward, total/leisure/business adoption is set to increase 3- has slowed in the US and Europe, as the percent of travelers who 4bp from /%/% in 7 to 9%// in 8, used over the last months increased by only ~33bp to respectively. (see Exhibit ). Two points stand out. First, this is a notable deceleration in the adoption curve, as user penetration increased by Exhibit ~8bp in 6. Second, adoption expectations of 9% for 8 are Our AlphaWise survey shows similar trends for business and leisure roughly flat to last year's expectations of for 7, so there is a travelers chance adoption could continue to undershoot estimates. Said Use another way, we're a year behind where we'd thought we'd be...and, were our prior optimistic estimates. % +6bp +543bp % 9% % +79bp Business Travelers 8E +6bp 7E in effect, 's adoption curve is flattening faster than what Leisure Travelers % % % Source: AlphaWise, organ Stanley Research 6 7A Business Travelers Leisure Travelers 8E 7E 7A 5 %

7 ...as Awareness Limits, Privacy and Safety Issues Hold Back Pace Note too that not only is the proportion of travelers not using of because of privacy and safety concerns on the rise, but the absolute number of people concerned about these issues also continues to But why is adoption slowing? Part of the deceleration is argu- increase (rising %/ in 7; see Exhibit 5 ). This is surprising ably due to slowing incremental awareness as awareness is and potentially troubling for 's growth potential because now at across the US, UK, Germany, and France online con- typically consumers become more comfortable with emerging sumers that we surveyed (rising only 8bp y/y this year compared technologies as awareness/testing/adoption grow. This doesn't to,bp last year; see Exhibit 3 ). And with awareness appear to be happening for, even as the company has begun now at, awareness-driven adoption can only go so far using marketing/grass roots initiatives to improve its perception. among this demographic in these countries. This, in our view, could speak to two potential truths: ) how (and the sharing lodging product) could be more niche than previously expected; and ) how the many opponents of may While awareness continues to rise, our survey suggests it has have gained some traction in combatting its growth. Looking reached a level where there is limited incremental upside ahead, we expect may need to increase its advertising and % of Respondents Who Have Heard of Exhibit 3: % branding campaigns to improve its perception...and the impact of 8 9% growth curve. these investments is likely to be critical to the company's forward 79% 5 5% 4 Exhibit 5 The absolute number of survey respondents concerned about privacy and safety has risen in each survey Primary Barriers for (Among Non- Users) % + %,6,49 +4 %,4 Overall France 5 6 US 7 UK +% Germany,,47 +49%,46 Source: AlphaWise, organ Stanley Research 949, And privacy and safety remain material and growing barriers to adoption. As shown below, the percent of non users who cite Privacy Concerns privacy as a concern grew ~7bp to 3. Safety also remains a hindrance to growth, with of non users (up ~4bp) listing Safety Concerns 5 Privacy & Safety Combined 6 7 Source: AlphaWise, organ Stanley Research safety as their reason to not use the platform. Exhibit4 Given the Slowing Adoption, We Lower Our Forward Privacy and Safety concerns are the top reasons why survey respon- Growth Expectations dents have not used (among those that have heard of ) Primary Barriers for (Among Non- Users) The slowing user adoption laid out above causes us to decrease our forward user forecast and room night forecast. Given % 9% the results of the last two surveys, we had previously modeled usage -,bp +653bp +36bp +33bp -48bp +86bp +58bp flattening, our forecast for user adoption in '8/'9/' falls to growing in a relatively straight line. Now that it appears the curve is 59% 39% // from a prior /3/4. We now expect to 3 grow to of lodging demand (across the US and Europe) by, 9% % % 9% Lack of Availability Lack of Amenities % Never heard of Privacy Concerns Safety Concerns 5 Source: AlphaWise, organ Stanley Research ORGAN STANLEY RESEARCH Uncertain Logistics 6 Other reason 7 vs. our previous estimate of 9% (see Exhibit 7 ). If, as we are currently modeling, 's share of total lodging demand begins to plateau, it may not pose as large a threat to traditional hotels as we had previously thought. 7

8 3 Source: % % e 9e STR, KG, organ Stanley Estimates (based on AlphaWise data) % % 8e Current Est e 9e 8e 7e 6 5 Prior Est. 7e Current % of US/Europe hotel demand 5 % 5 Prior % of US/Europe hotel demand 5 48 Room Nights (in millions) Est. Users (in millions) 6 6 Source: Prior Room Nights Est. Current Room Nights Est Exhibit 7...and now expecting to grow to of hotel demand (across the US and Europe) by % of US/EU hotel demand We are lowering our user forecast by /39%/4 in '8/'9/ '... 9 Exhibit6 STR, KG, organ Stanley Estimates (based on AlphaWise data) In all, we expect to Generate ~55mn Room Nights in 8 in the US/Europe, up ~ Y/Y Exhibit 8 below lays out our summary model, where we show how the Exhibit 8: user penetration combined Base Case: We estimate that will generate ~55mn room nights in the US and Europe in 8, with an assumed average 4.3 nights up Y/Y booked per user per booking, times US & Europe Room Night Calculation (mn) Addressable User arket Penetration stayed at per user per year, and % 7e 6 8e % users per booking (all from our survey work detailed in today's note), we arrive at a total estimated ~55mn room nights in 8 (across the US and Europe). We note that this is just our base case, as our bull case (a worst outcome Users (x) Room Nights/User/Booking (x) Times Stayed at /User/Year ( ) Number of Users/Booking Total Room Nights Y/Y % Growth Y/Y Abs. Growth Source: AlphaWise, organ Stanley Research Estimates (based on AlphaWise Data), organ Stanley Research for hotels) assumes this survey was just a blip and that adoption re-accelerates and travelers use more fre- Exhibit 9: quently. This leads to ~mn 8 room We believe that could grow to as many as ~mn room nights in 8 or as few as nights, as room night growth increases on ~4mn an absolute basis and awareness/accept- (mn, excluding per user data) ance rise...and is able to somewhat Scenario Analysis Se Se Se Bear Base Bull overcome its privacy and safety chal- Addressable User arket 7 6 User Penetration lenges. 8e % 3 Room Nights per User/Booking Our bear case assumes that penetration Times Stayed at per User/Year growth in the US/Europe continues to Number of Users/Booking Total Room Nights slow dramatically and frequency declines. We model ~4mn 8 room nights, as adoption slows and future penetration comes from less frequent travelers, since veterans will have already made up their minds on. 8 Source: AlphaWise, organ Stanley Research Estimates (based on AlphaWise Data), organ Stanley Research

9 Other odel Adjustments. It is notable in the above charts Exhibit : that we are making other adjustments to our model for impact on hotels' RevPAR: Our updated estimates suggests less increased precision. First, we've adjusted historical user figures to of a drag from... account for y/y population growth across the US/Europe and are now.% using an online penetration rate of online leisure/unmanaged busi- -.% ness travel vs. leisure only. Second, we are now using the combined -.% leisure/business user penetration rate from our survey vs. lei- -. sure only previously to provide a more complete view of user -. adoption. As a result, our '5/'6 user figures decrease by given a lower addressable market and higher user penetration. How- -. ever, our estimated '5/'6 room nights increase - given -. the higher penetration numbers New Prior % -.9% -.% ) We are raising our 8 / 9 traditional hotel RevPAR esti- Source: organ Stanley Research mates, which drives estimate and price target revisions across our Lodging coverage. With our AlphaWise survey suggesting slowing Exhibit : penetration, our model implies...as a result, we are raising our US RevPAR forecasts, and now expect 8 growth of. (up from less impact on traditional hotels our prior estimate of.%) and establish a. 9 estimate RevPAR in the US and Europe. We now 3. 3.% forecast 8e US RevPAR growth of -... (up from prior.%) which now contemplates a 5bps headwind from which incorporates an additional 4bps headwind and a stable economic. -. (prior 8bps). For 9, we expect US RevPAR growth of., -.%.% RevPAR Supply OTA Pricing Pressure Corp Travel Rates acro 7E RevPAR Supply OTA Pricing Pressure acro 8E RevPAR OTA 9E Pricing RevPAR Pressure backdrop. -. Source: organ Stanley Research; STR Summary of Lodging Estimate / Price Target / Risk-Reward Exhibit : Changes. We are raising our 8 / 9 EBITDA estimates by. Lodging: New vs. Old Price Targets /. on average. We have also rolled forward our price targets to bear case multiples for our stocks, which results in the risk-reward Ticker Lodging C-Corps HLT AR LQ H STAY CHH skew for our coverage going from neutral to positive. Note we con- C-Corps Avg tinue to prefer the asset-light C-Corps over the REITs given the bene- Lodging REITs HST LHO XHR DRH SHO YE8 (based on 9e EBITDA) given increased visibility / confidence in a stable environment, which results in higher price targets and on average, and upside to our price targets including dividends. In addition, as we see as less of a structural threat, we raise the fits of unit growth in a low RevPAR growth environment, and their relative exposures suggesting much lower supply growth. Rating OW OW EW EW EW EW Price /8/7 % 9% % -% % - % - REITs Avg - -% Covered Lodging Avg % $ $ $9 7 5 $ Upside (%) New PT + Div % % EW EW EW UW UW $ Price Target Old New $9 7 5 Source: Company data, Thomson Reuters, organ Stanley Research ORGAN STANLEY RESEARCH 9

10 3) Slowing Adoption Speaks to OTAs' Competitive oats Exhibit 3: The OTAs are the only channel showing substantive growth, other than Channels Used to Book LeisureTravel (Among Past onths Travelers) Our survey data show that OTA channel usage is still on the rise, up ~6bp y/y (see Exhibit 3 ) as this appears to be the only channel showing sub- stantive growth other than. In our view, this speaks to the strength of the OTA model and 7 +59bp +bp p -54bp -69bp +6bp user proposition, even as continues to 3 grow its use base % % % % % If Anything, OTAs Are Becoming Larger Threats to as Potential Online Travel One-Stop- % Online travel agency Hotel website Called or walked into hotel Travel agent Vacation rental site Shops 5 Traffic matters to any e-commerce business, and 6 7 Source: AlphaWise, organ Stanley Research the OTAs are already expert paid search spenders and traffic acquirers (expected to spend ~$9bn on advertising in 7). And over the long-term, we see the sites that offer consumers the most selection across hotels and alternative accom- Exhibit 4: ~6 of travelers visit -3 sites when booking their online travel and 4 visit - Number of Travel Websites Visited While Planning Leisure Trip modations as likely being the best positioned for repeat traffic and strong brand equity. This too plays to OTAs' advantage, as they are already working to build on their leading hotel offerings to show consumers hotels and % % alternative accommodations. In effect, over the long-term, the OTAs are becoming growing threats to 's long-term growth opportunity. % 9% % % - ore than % None Source: AlphaWise, organ Stanley Research

11 Exhibit 5 PCLN's Booking.com allows travelers to comparison shop between hotels and alternative accommodations... Source: Booking.com, organ Stanley Research Q: Q:6 4Q: Q: Q: Q: Q: Q:4-4Q:7E Hotel CAGR: Vacation Rental CAGR: 5% 9 69, # of Booking.com Vacation Rental Properties 4Q:7E 3Q:7 Q:7 3Q:6 Q:5 4Q:4 Booking.com Property Count (in thousands) Exhibit 6...and PCLN has been building its alternative accommodation inventory for many years # of Booking.com Hotel Properties Source: AlphaWise, organ Stanley Research ORGAN STANLEY RESEARCH

12 Key Debate #: How Quickly Is User Adoption Growing? New Survey Data Speak to Slowing Adoption in the US arket View: Continued Strong Adoption: incremental user adoption continues to be very strong in US / Europe across leisure and business travelers. Our View: Slowing Adoption Due to Growing Privacy and Safety Concerns Our AlphaWise survey data indicate that in the past year, adoption growth has slowed, as the percent of travelers who used over the last months increased by only ~33bp to, a notable deceleration in the adoption curve, as user penetration increased by ~8bp in 6. Adoption expectations of 9% for 8 are roughly flat to last year's expectations for 7 of, so there is a chance adoption could continue to undershoot estimates. Awareness continues growing, but awareness can only drive so much adoption, as privacy and safety remain material and growing barriers to adoption. Given these trends, we are reducing our forward forecast, now expecting the company to reach only of US/Europe room nights, vs 9% previously. and Europe... According to our survey, in the past year, adoption has slowed in the US, UK, Germany, and France, as the percent of travelers who used over the last months increased by only ~33bp to (see Exhibit 7 ). Two points stand out. First, this is a notable deceleration in the adoption curve, as user penetration increased by ~8bp in 6. Second, adoption expectations of 9% for 8 are roughly flat to last year's expectations for 7 of, so there is a chance adoption could continue to undershoot estimates. In effect, the adoption curve in the US and Europe is flattening faster than expected. Exhibit 7 Our AlphaWise survey shows of total travelers used in the last months...lower vs. expectation of in last year's survey, indicating respondents may be overestimating future use Use (Total Travelers) 3 +39bp 9% +788bp % % % % Where We Could Be Wrong: 5 Source: 6 Total Travelers 7A 7E 8E AlphaWise, organ Stanley Research could successfully work around the large/growing privacy and safety concerns and drive faster user This trend holds both for leisure and business travelers, with actual adoption. adoption over the last months roughly 5bp below last year's forward expectation for both types of travelers (see Exhibit 8 ). Leisure travel adoption increased from 9% in 6 to % in 7 (vs. expected adoption of ); and business usage grew from in 6 to % in 7 (vs. expected adoption of ). Looking forward, total/leisure/business usage is expected to increase 34bp from /%/% in 7 to 9%// in 8, respectively.

13 Exhibit 8 One caveat is our survey is internet-based, suggesting that respon- Our AlphaWise survey shows similar trends for business and leisure travelers dents are likely younger and more tech savvy, so overall penetration of total potential travelers may be lower. Use Leisure Travelers +6bp +79bp Business Travelers +6bp % % +543bp % 9% % % % Source: 8E 7E 7A 6 Business Travelers 5 8E 7E 7A 6 5 % Leisure Travelers Exhibit : While awareness continues to rise... % of Respondents Who Have Heard of % 8 9% % 7 5 5% 4 % % AlphaWise, organ Stanley Research % Overall While adoption appears to be slowing, it is notable that user fre- France 5 6 US 7 UK Germany Source: AlphaWise, organ Stanley Research quency remained roughly constant year-over-year, as the average user used the service ~3X both this year and the last two The top reasons cited by people in our survey for not using are years. lack of awareness (), privacy concerns (3), and safety concerns (). The year-over-year trends are notable (and a potential long-term challenge), as 3/ of respondents now view privacy/ Exhibit 9 user frequency remained roughly unchanged Y/Y safety as the main barriers to using...up from /9% last Number of Times Used in Past onths (Among Users of ) year. Weighted average 5: times per year 6: 3. times per year 7: times per year In our view, lack of awareness is a relatively simple problem to solve (with increased advertising and availability, etc.), but the privacy and 9% safety concerns may be bigger challenges to long-term growth given 3 s social sharing model. This is particularly important for corporate users, whose travel managers and companies need to have % % safety and regulations in mind. Privacy and safety concerns con- % ore than 7 Source: AlphaWise, organ Stanley Research Respondent Awareness at (up from ), but Privacy and Safety Issues Hold Back Pace of In aggregate, consumer awareness is up to from last year, ranging from 7 to 8 (with Germany the lowest and France the highest) and rising y/y in every country we surveyed. As with last tinue to grow as awareness of increases, which indicates to us that they may pose a significant barrier to widespread adoption. Exhibit...Privacy and Safety concerns are the top reasons why survey respondents have not used (among those that have heard of ) Primary Barriers for (Among Non- Users) % 9% -,bp year's survey, rising awareness seems to be driving adoption. We see that awareness rose by ~8bp y/y and usage grew ~33bp. However, both of these growth rates are ~ of what we saw last high awareness () may make incremental gains more difficult. +36bp +33bp -48bp +86bp +58bp 59% year. We expect awareness growth to slow, since 's relatively +653bp 39% 3 9% % % 9% Lack of Availability Lack of Amenities % Never heard of Privacy Concerns Safety Concerns 5 Uncertain Logistics 6 Other reason 7 Source: AlphaWise, organ Stanley Research ORGAN STANLEY RESEARCH 3

14 Note too that while the proportion of travelers avoiding Given the Slowing Adoption, We Are Lowering Our Forward because of privacy and safety concerns is rising, the absolute number Growth Expectations of people concerned about these issues also continues to increase (see Exhibit ). This again is surprising and potentially troubling for The slowing user adoption laid out above cause us to decrease 's growth potential because typically consumers become our forward user forecast and room night forecast. In all, our forecast more comfortable with emerging technologies as awareness/ for user adoption in '8/'9/' falls to // from a prior testing/adoption grow. This doesn't appear to be happening for /3/4, reflecting a flattening adoption curve versus rela- even as the company has begun using marketing/grass roots tively straight growth assumption last year. As such, we now expect initiatives to improve its perception. This, in our view, speaks to how to grow to of lodging demand (across the US and Europe) (and the sharing lodging product) could be more niche than by, vs. our previous estimate of 9%. If, as we are currently previously expected. modeling, 's share of total lodging demand begins to plateau, it may not pose as large a threat to traditional hotels as we had previously thought. Exhibit The absolute number of survey respondents concerned about privacy and safety has risen in each survey Primary Barriers for (Among Non- Users) Exhibit 4 +,6, %, Safety Concerns Privacy & Safety Combined bp +4bp +86bp % % % % Source: % % Less expensive Location Authentic experience Easy to use site/app 5 Source: AlphaWise, organ Stanley Research 6 Own kitchen 7 Uniqueness of unit Can accommodate large party Other reason % % % +47bp -35bp % STR, KG, organ Stanley Estimates (based on AlphaWise data) e 5 +6bp 8e 5-48bp % +46bp 5 9% Current % of US/Europe hotel demand 5 64 % Prior % of US/Europe hotel demand 74 Factors of Importance When Using (Among Past/Next onths Users) Room Nights (in millions) Price is the most important factor leading to use 9 Current Room Nights Est. 3 Exhibit 3: Prior Room Nights Est. % of US/EU hotel demand 35 is.7x larger than "location" (3) and "authentic experience" (3). 7e them to use or plan to use the service (up from 5 last year). This Exhibit 5...now expecting to grow to of hotel demand (across the US and Europe) by 7 that "price" was one of the three most important factors that led In all, 5 of respondents who either used or plan to use said Current Est. STR, KG, organ Stanley Estimates (based on AlphaWise data) 6 Price Still the ost Important Factor in Driving Use 9e 6 5 Prior Est. Source: e Source: AlphaWise, organ Stanley Research e Privacy Concerns e e, 73 8,47 +49%, Est. Users (in millions),4 We are lowering our user forecast by /39%/4 in '8/'9/ '...

15 We Expect to Generate ~55mn Room Nights in 8 in the US/Europe, up ~ Y/Y Combining the ~mn addressable user market, user penetration, average 4.3 nights booked per user per booking, times stayed at per user per year, and 3. users per booking, we arrive at a total estimated ~55mn room nights in 8 (across the US and Europe, see Exhibit 6 ). This room night usage is up only Y/Y, compared to the 59% Y/Y increase we see in 6, driven by the significant reduction in our estimated penetration growth rate. While there are reports suggesting that overall revenue is growing much faster (7-), this is likely being driven by new and under-penetrated geographies, like China; the new trips product; and 's purchase earlier this year of Luxury Retreats. Exhibit 6: Base Case: We estimate that will generate ~55mn room nights in the US and Europe in 8, up Y/Y US & Europe Room Night Calculation (mn) Addressable User arket Penetration Users (x) Room Nights/User/Booking (x) Times Stayed at /User/Year ( ) Number of Users/Booking Total Room Nights Y/Y % Growth Y/Y Abs. Growth % 7e 6 8e % Source: AlphaWise, organ Stanley Research We note that this is just our base case, as our bull case (a worse outcome for hotels) assumes that adoption re-accelerates and travelers use more frequently. This leads to ~mn 8 room nights, as room night growth increases on an absolute basis and awareness/acceptance rise...and is able to somewhat overcome its privacy and safety challenges. Our bear case assumes that penetration growth decelerates more significantly and frequency declines. We model ~4mn 8 room nights, as adoption slows and future penetration comes from less frequent travelers, since veterans will have already made up their minds on. Exhibit 7: We believe that could grow to as many as ~mn room nights in 8 or as few as ~4mn (mn, excluding per user data) Scenario Analysis Se Se Se Bear Base Bull Addressable User arket 7 6 User Penetration % 3 Room Nights per User/Booking Times Stayed at per User/Year Number of Users/Booking Total Room Nights e Source: AlphaWise, organ Stanley Research ORGAN STANLEY RESEARCH 5

16 Key Debate #: How Big of a Competitive Threat Is to Hotels? We continue to believe that does represent a headwind for the broader US Lodging industry, though with penetration rates slowing, we see the structural threat as lower than we previously thought. As previously discussed, our AlphaWise survey suggests that penetration growth is slowing, and below our prior expectations (see Exhibit ). As a result, our model implies is having less of an impact on traditional hotels RevPAR in the US and Europe. We now forecast 8e US RevPAR growth of. (up from prior.%) which now contemplates a 5bps headwind from (prior 8bps). For 9, we expect RevPAR growth of., which incorporates an additional 4bps headwind and a stable economic backdrop. Exhibit 8: Exhibit 9: impact on hotels' RevPAR: Our updated esti- We are raising our US RevPAR forecasts, now expecting 8 growth of. (up from our mates suggest less of a drag prior estimate of.%) 3..% 7 -.% 3.% % % -.% New -. Prior % -. 6 RevPAR Supply -.9% OTA Pricing Pressure Corp Travel Rates acro 7E RevPAR Supply OTA Pricing Pressure acro 8E RevPAR OTA 9E Pricing RevPAR Pressure -.% -. Source: organ Stanley Research Source: organ Stanley Research; STR We base our US RevPAR forecasts on a build-up of the macro and our price targets only increase for the Lodging C-Corps. This is structural factors that drive growth. In terms of structural factors, because in a low RevPAR environment, the REITs will see flat to in addition to, we also see traditional supply growth, pricing declining EBITDA given cost inflation, while they use the majority of pressure from traditional hotels discounting rates to compete with their cash flow for dividends. On the other hand, the C-Corps' unit OTAs, and corporate rate negotiations as having an influence on growth and free cash flow imply rising values on a stable multiple. future RevPAR growth. In terms of macro factors, we use organ Stanley forecasts for business fixed investment spending and Exhibit 3: employment growth as factors to drive our forecasts. Net-net, this New vs. Old Price Targets implies moderately slowing RevPAR into 9. Our price targets for our Lodging stocks increase by on average, resulting in average upside. We have adjusted our models to reflect our new more tempered impact expectations, and shifted our valuations to be based off of 9 EBITDA rather than 8. We do not change our base case valuation multiples as we believe a stable to improving economic environment should support the higher multiples we have been using. As a result, on average, our price targets increase by. However, while we have raised our estimates for almost all the stocks in our coverage (ex Extended Stay where it appears the threat to this segment of the industry is higher, and for Sunstone as we scrubbed our numbers), 6 Ticker Lodging C-Corps HLT AR LQ H STAY CHH Rating OW OW EW EW EW EW Price /8/7 $ Price Target Old New $ $ Upside (%) New PT + Div % % % 9% % -% % - % - REITs Avg - -% Covered Lodging Avg % C-Corps Avg Lodging REITs HST LHO XHR DRH SHO EW EW EW UW UW $ $9 7 5 Source: Company data, Thomson Reuters, organ Stanley Research $9 7 5

17 We have adjusted multiples in our bear cases, and see the current risk-reward skew as more attractive. With our survey results suggesting that is less of a Exhibit 3: With as less of a structural threat, we have adjusted our bull and bear cases, which has resulted in a slightly positive skew for our coverage universe structural threat to hotels than we had previously been concerned about, we raise our bear case multiples. We 9% now reflect the higher quality Lodging C-Corps (Hilton, Hyatt, arriott, Choice) not dropping below a x mul- tiple even if our bear case scenario of a mild US recession comes to fruition, which would imply one year of down % Current price RevPAR growth. A ~.5x multiple would be in line BULL % a potential recession in early 6, and given the other - lodging stocks now more closely mirror AR in terms of - 3 % -% with where AR traded to when the market discounted 3 3 BASE - Underweight Equal-weight Overweight - - PT - -% % % -3 LHO XHR DRH SHO BEAR HLT AR LQ H STAY CHH HST asset intensity, we see it as an appropriate bear case multiple. The result is that the bull/bear skews are now Source: organ Stanley Research; ThomsonReuters weighted to the upside, versus prior the skew was even. Exhibit 3: 's users (across leisure and business travelers) replace a variety of other accommodations, but mainly (and increasingly) hotels While the overall impact of on our RevPAR forecasts is lower than prior expectations, we still see Accommodation Alternatives that Replaced (Among Total Users Past onths) impacting hotels occupancy, especially as using as a substitute for traditional hotels continues to inch up to around the range. Our survey results suggest that 5% of users replaced stays at traditional hotels, an increase of ~6bps from 49% in 6, though at a much slower rate than the ~8bps increase from 5 to 6. Survey respon- % dents expect this cannibalization rate to increase further % to 5 in the next months. Interestingly, the replace- % +6bp +bp +56bp -9bp -39bp -33bp -5bp +56bp 49%5% % %% %% 9% %% 9% Traditional hotel ment of extended stay hotels increased by 6bps to -48bp Bed & Breakfast Friends/family Extended stay hotel in 7, after declining 3bps from 5 to 6. 5 Vacation rental site 6 Other Vacation Rentals % Other I would not have taken this trip if not for Corporate apartment 7 Source: AlphaWise, organ Stanley Research We note that commentary from hoteliers remains mixed on the impact to their businesses from. Large-scale brand companies have generally downplayed the impact (e.g. arriott has said / homesharing has been less impactful than one would have Exhibit 33: Looking ahead, 5 of users (leisure and business) expect to continue to shift their accommodation stays away from hotels over the next months Accommodation Alternatives that Likely to Replace N (Among Total Users Next onths) expected over the last few years, and even less so today, and Hilton has said there has not been a material impact). On the other hand, some of the Lodging REITs (which are concentrated in gateway cities) have suggested that the rise in short-term rentals has magnified the impact of supply growth for the industry, with the impact espe- +473bp -7p +7bp +34bp -bp -4bp +6bp +86bp +8bp 5 49% % % 9% cially felt during compression nights (see Exhibit 4 ). % % % % 9% % % Traditional hotel Bed & Breakfast Friends/family Extended stay Other Vacation Vacation rental hotel Rentals site 5 6 Corporate apartment Other Would not take trip if not for 7 Source: AlphaWise, organ Stanley Research ORGAN STANLEY RESEARCH 7

18 Our 7 survey suggested a modest shift in terms of (up from % in the prior years). This is supportive of the view that users shortening their length of stay, but users continue trips don't always cannibalize traditional hotels even with an to average ~4 nights per stay, longer than the traditional hotel user, as our survey found that 4 of traditional hotel book- stay. The weighted average length of stay among users in ings are for stays of - nights, while only of bookings are. 7 was 4.3 nights, down slightly from 4.5 nights in 5/6, with staying night (up from last year) and staying two nights Exhibit 34 users average ~4 nights per stay % of Survey Respondents Weighted average 5: 4.5 nights/booking 6: 4.5 nights/booking 7: 4.3 nights/booking Exhibit 35:...and is less focused on single night stays, which is where hote5 5% liers are focused with their corporate demand 7: Average Number of Nights per Booking (Among Users) 7 %% % % 3 % 7 Traditional Hotel % % % % night nights 3-5 nights 6- nights ore than nights Nights Stayed per Booking When Using Source: AlphaWise, organ Stanley Research % % night nights 3-5 nights 6- nights ore than nights Source: AlphaWise, organ Stanley Research Exhibit 36: Travelers who Take 5- Trips / Year Allocated the ost of Their Trips to in the LT. It Appears Road Warriors Concentrate Their Trips ore on Traditional Hotels Traveler Profile % of Total Total Light Travelers (-4 Trips / Year) 49% Frequent Travelers (5- Trips / Year) Road Warriors (+ Trips / Year) Used for Leisure Travel LT Used for Business Travel LT % % 3 Number of Times Used LT (Among Users of ) : % of Total Trips % 9% % % % 3 9% % % 3 9% % % 9% 3 9% % 4 % % % % 5% % % 49% % % Leisure Share (based on most recent trip) HomeAway/VRBO/Other Hotel website Called or walked into hotel Travel agent Online travel agency Other channel Business Share (based on most recent trip) HomeAway/VRBO/Other Traditional hotel Extended stay hotel Corporate apartment Friends/family Bed & Breakfast Other channel Source: organ Stanley AlphaWise 8

19 Breaking Down the Impact of on Traditional Hotels As we have outlined prior, our base Exhibit 37: case assumes that room Base Case: We estimate that will generate ~55mn room nights in the US and Europe in 8, up nights in the US and Europe Y/Y increase in 7 and in US & Europe Room Night Calculation (mn) Addressable User arket Penetration 8, reaching ~55mn. For traditional hotels, our forecasts would have called for a blended occupancy in the US and Europe of 68.% in 6 to rise to 69.% in 7 and 7.% in 8 if didn't have an impact. Users (x) Room Nights/User/Booking (x) Times Stayed at /User/Year ( ) Number of Users/Booking Total Room Nights Y/Y % Growth Y/Y Abs. Growth % 7e 6 8e % Source: AlphaWise, organ Stanley Research Assuming still record and growing occupancy, we assumed room rates would grow. in 7 and.% in 8, slowed only by discounted pricing by hotel brands to combat OTAs. Exhibit 38: Traditional Hotel odel Assuming Constant acro and No for Forecast Periods Traditional Hotel odel Before Impact Se Traditional Hotel Demand (Room Nights, m) % Growth Europe hotel rooms % Growth US hotel rooms % Growth Se Traditional Hotel Supply (Room Nights, m) % Growth 4,696 5, ,95.9% 6,83.% 6.3.9% ,44.% 7E,98 3.% 6.38.% 5.3.% 4,4. 8E,995 3.% 6.45.% 5.4.% 4,67. 9E 3,85 3.% % 4,34. E 3,77 3.% 6.68.% 5.46.% 4,49..% Implied Occupancy (ex impact for forecast years) % Growth % 67..% 68.%. 69.%.% 7.%.% 7.%. 7.. Expected Rate Growth at Estimated Occupancy Levels. 6.%.9%..%.9%. Se RevPAR Growth (ex impact for forecast years) % ,57 Source: organ Stanley Research; STR; KG Exhibit 39: drag to annual occupancy over the next two years. We assume Traditional hotel demand lost to nights will be a ~5bps annual that penetration increases to 7. in 8, with this demand drag to occupancy survey), implying a ~5bps annual drag on occupancy. However, it's important to note that our estimates do assume that occupancy in aggregate continues to increase, assuming stable ~ annual demand growth, only partially offset by.5-% traditional hotel supply growth and the demand cannibalization. Increasing occupancy is important because it's when occupancy starts to decline when hoteliers start to be more aggressive on room rates. 4,5 3,,696,769,83,898,974 7.%, %,5 69% 6 68.%, 69.%, , Hotel Demand Hotel Supply Hotel Dccupancy Source: organ Stanley Estimates (based on AlphaWise data) ORGAN STANLEY RESEARCH 4,67 4,4 4,44 4,95 4,57 4, Hotel Occupancy cannibalizing traditional hotels at a 5 rate (also per our AlphaWise US and Europe Room nights (mn) Our new base case forecasts suggest that will be a ~5bps 8 Hotel occupancy ex 9

20 As a result of the occupancy drag Exhibit 4: caused by, we assume a sim- The average change in rate growth compared to the change in occupancy has been.x over the past ilar impact on hotel pricing power. 8 years We assume that the occupancy declines being caused by will drive lower overall pricing growth for the industry. While the relationship between occupancy changes and pricing growth is volatile, the historical average over the past 8 years has been an even.x. So to estimate the impact Occ. Growth Rate Chg Growth ultiplier Occ. Growth Rate Chg Growth ultiplier % -.% 3.% % 5.9% % -.%.6x.x % %.x x x...x x x 5. -.% 7.4x 6.% -..5x.9x.x -.x.x -.x -9.8x.8x.3x Source: organ Stanley Research; STR on pricing from, we assume a x multiplier effect on the -driven occupancy decline. Net-net, we forecast a 5bps drag on our US RevPAR growth forecast in both 7/8 from, with a slightly lesser Exhibit 4: Build-up of our base case estimate of 's impact on traditional hotels US and Europe Traditional Hotel odel Impact Hotel Supply (Rooms m) (A) % Growth 7e.5. 8e.7. 9e.9. e..% Hotel Occupancy (B) 69.% 7.% 7.% 7. Hotel Demand (Room Nights, m) (A x B x 365 = C) % Growth,98 3.%,995 3.% 3,85 3.% 3,77 3.% Estimated User Penetration of Online (D) Change (bps) 3bps 5bps bps 5bps 4bps drag in 9. Top 5 US markets: Where the risk is higher. While our survey tempers our concerns for the overall hotel industry, we still see the top 5 Online penetration growth (E) 4.% 4.% 3.% 3.% markets in the US as at risk, as illus- Room Nights (mn) (growth in D and E = F) as % of Hotel Room Nights % % 5% Annual impact on hotel occupancy (H / C = I).%.%.%.% Commensurate impact on ADR growth (J).%..%.% during these peak demand periods Annual impact on RevPAR growth (I + J).... where operators have the ability to Source: organ Stanley Research; STR; KG trated through compression nights. We define a compression night as any % Cannibalizing Traditional Hotel Demand (G) night with >9 occupancy. We see Incremental room nights coming from hotels (G x change in F = H) compression nights as a good proxy for the impact of supply growth, including alternative accommodations, as it's push rate the most, unless there are substitutes. We saw a drop in compression nights in top 5 markets in the US result than in the US, but note that Europe's performance comes off in 6, the first decline since 9. YTD the number of compression a weak base as some markets suffered after last year's terror attacks nights are currently tracking down an additional. While com- (see report here). pression nights in San Francisco (which contributes ~% of annual compression nights) are down as a result of its largest conven- It's important to note that traditional hotel supply growth, tion space, the oscone Center, being closed, even ex San Francisco, which is being concentrated in gateway cities, is also likely compression nights are tracking down % this year. On the one having an impact on compression nights. According to Lodging hand, this highlights the fact that rising supply growth and increasing Econometrics, supply growth in the top 5 US markets was. in usage of alternative accommodations continues to negatively impact 7, significantly higher than the.% (according to STR) seen hotels during peak periods. However, the decline ex-sf also suggests across the broader US. In fact, the headwind will get worse in 8, a lesser headwind than last year, consistent with our survey results. with top 5 markets supply growth accelerating 5bps to 3. com- In Europe, the number of compression nights in the 6 European cities pared to the broader US increasing just bps to.%. This will likely we looked at increased by 3 in June YTD, a more encouraging lead to the top 5 markets seeing further pressure on the number of

21 compression nights in 8. This fur- Exhibit 4: ther drives our preference for asset- Number of Compression Nights in Top 5 US arkets Shows Declines the Past Years 4 -, -% ex SF YTD 6 YTD Diamondrock and Sunstone heavy, gateway city focused REITs, like like Hilton and arriott, over asset- # of Compression Nights in Top 5 kts light, geographically diverse C-Corps Source: STR; organ Stanley Research Exhibit 43: YTD Declines in Top 5 US arkets' Compression Nights Have Been Relatively Broad-Based # of Compression Nights Anaheim-Santa Ana, CA Atlanta, GA Boston, A Chicago, IL Dallas, TX Denver, CO Detroit, I Houston, TX Los Angeles-Long Beach, CA iami-hialeah, FL inneapolis-st Paul, N-WI Nashville, TN New Orleans, LA New York, NY Norfolk-Virginia Beach, VA Oahu Island, HI Orlando, FL Philadelphia, PA-NJ Phoenix, AZ San Diego, CA San Francisco/San ateo, CA Seattle, WA St Louis, O-IL Tampa-St Petersburg, FL Washington, DC-D-VA Total YTD YTD Source: STR; organ Stanley Research ORGAN STANLEY RESEARCH

22 Exhibit 44: The number of compression nights has increased this year in the European cities we looked at, although still below 5 for some Note: Compression nights refer to the number of nights that occupancy in a market is higher than 9%. ADR = Average daily room rate. Source: STR, organ Stanley Research.

23 Key Debate #3: How Big a Competitive Threat Is to OTAs? Slowing Adoption Speaks to How Still Not a arket View: Unclear, as the market debates whether is a bigger threat (a new, rapidly-growing platform offering alternative inventory), neutral, or a friend (negatively impacting the hotels' business). Our View: Threat to the OTAs... With another year of survey data (and the OTAs' evolving strategies), we continue to be of the view that is not a material threat to the OTAs: First, our survey data show that the OTA channel is proving to be relatively resilient, with usage up ~6bp y/y (see Exhibit 45 ). OTAs With three years of data points, we are even less appear to be the only channel showing substantive growth other concerned that is a material competitive threat than, which rose ~6bp y/y. In our view, this speaks to the to the OTAs. First, our survey data show the OTA channel opportunity for both the OTAs and to grow in a large and fast- has been relatively more resilient than the hotel channel growing travel space. over the past three years. Second, if hotels continue to lose share to, they could become more dependent Exhibit 45: on OTAs for traffic which will be positive for OTA take The OTAs are the only channel showing substantive growth, other than rates and profitability. Third, the major global OTAs have been building out their inventory of "non-hotel" Channels Used to Book LeisureTravel (Among Past onths Travelers) accommodations (bed & breakfasts, chalets, apartments, 7 etc). Fourth, we believe the OTAs' search marketing 6 expertise and the sheer dollar spent on advertising (~ 5 $9bn in direct advertising annually) are barriers for 4 to surmount. If anything, the OTAs are becoming a larger threat to. +59bp +bp p -54bp -69bp +6bp % % % % % Where We Could Be Wrong: % Online travel agency Hotel website Called or walked into hotel If ever decides to allow hotels to list properties on their site a likely positive for 's user conversion 5 6 Travel agent Vacation rental site 7 Source: AlphaWise, organ Stanley Research and monetization it could create a more comprehensive, faster growing, lower cost OTA competitor. currently charges properties on transactions...while OTAs charge hotels -. Even if offered the hotels a commission rate that is triple its standard rate, it would be positive for hotels, likely leading to lower average customer acquisition costs and it would give the hotels more leverage against the OTAs. ORGAN STANLEY RESEARCH Second, while incremental adoption is slowing, demand continues to cannibalize traditional branded hotel business...as our survey data show that 5% of users replaced a traditional hotel with, up ~6bp y/y (see Exhibit 46 ). If hotels continue to lose share to, they could become more dependent on OTAs for online traffic which would be positive for OTA room night growth, take rates and profitability. 3

24 Exhibit 46: 's users (across leisure and business travelers) replace a variety of other accommodations, but mainly (and increasingly) hotels As the OTAs Are Becoming Larger Threats to as Potential Online Travel One-Stop-Shops Accommodation Alternatives that Replaced (Among Total Users Past onths) While consumers currently are generally using OTAs for hotels and for alternative accommodations we expect the lines to blur +6bp -48bp +bp +56bp -9bp -39bp -33bp -5bp +56bp 49%5% that offer the most inventory and selection across hotels and alternative accommodations. After all, our latest survey shows 9% that ~/3 of leisure travelers visit no more than 3 sites when booking %% %% 9% %% 9% % % their travel. The OTAs are investing/innovating to drive and ben % Other I would not have taken this trip if not for % Traditional hotel over the long term, as consumers are likely to favor the websites Bed & Breakfast Friends/family Extended stay hotel Source: AlphaWise, organ Stanley Research 5 Vacation rental site 6 Other Vacation Rentals Corporate apartment 7 efit from this trend. Exhibit 48: ~6 of travelers visit -3 sites when booking their online travel and 4 visit - Number of Travel Websites Visited While Planning Leisure Trip To this point, looking ahead, consumers expect to increasingly move away from booking direct with hotels as well (see Exhibit 47 ) as expected use of hotel websites and call/walk into hotels is expected to fall by 8bp/9bp y/y over the next months. OTAs use is expected to fall by ~bp, but this is materially less than the hotels' % % direct offerings. % Exhibit 47: The travelers in our survey indicated they plan to decrease their use of hotel websites and hotel call/walk-ins over the next months...while the OTA channel is expected to be relatively resilient 9% % % - ore than % None Source: AlphaWise, organ Stanley Research Channels Used to Book Leisure Travel; 7 9% % % % -bp -8bp +p -9bp +4bp +4bp First, the OTAs Are Integrating/Building Their Alternative Accommodation Inventory And while consumers continue to be interested in alternative accom3 modations (like ) it is notable that PCLN and EXPE are working % to build on their leading hotel offerings to show consumers hotels and alternative accommodations/vacation rentals. We see examples Online travel agency Hotel website Travel agent Past months Called or walked into hotel Next months Vacation rental site of this below as the leading OTAs are now showing "non-hotel" inventory (bed & breakfasts, chalets, apartments, etc.) Source: AlphaWise, organ Stanley Research As with any ecommerce site, selection matters to driving online user conversion, and we would argue the OTAs' ability to show alternative accommodations and hotels on their platform which is likely more convenient for shoppers gives them an advantage vs.. 4

25 Exhibit 49: Increasingly, Expedia is integrating alternative accommodation listings from HomeAway into its search results Source: Expedia, organ Stanley Research ORGAN STANLEY RESEARCH 5

26 Exhibit 5: PCLN's Booking.com allows travelers to comparison shop between hotels and alternative accommodations Source: Booking.com, organ Stanley Research 6

27 In fact, the number of vacation rental properties on PCLN's Exhibit 53 Booking.com now exceeds the number of hotels on the platform,...and we estimate that together, they spend ~X as much on marketing with vacation rentals growing at a 5% forward CAGR (vs. for as the top 5 largest hotel companies' marketing spend $7, hotels ). PCLN has been building its alternative accommodation inventory for $4, $3, ,53,7 3,775,4 $, 5,8 95 Expedia Priceline 6 75 Choice 5 $ $5, $, 4Q:4-4Q:7E Hotel CAGR: Vacation Rental CAGR: 5% 69, Hilton arriott arketing spend IHG 3 3 Accor Rewards program spend Source: Company Data, organ Stanley Research 's growing app user base and organic marketing have been large drivers of growth for the platform...but over the long-term we # of Booking.com Vacation Rental Properties Source: 4Q:7E 3Q:7 Q:7 Q:7 4Q:6 3Q:6 Q:6 Q:6 4Q:5 3Q:5 Q:5 Q:5 4Q:4 Booking.com Property Count (in thousands) many years in millions Exhibit 5 5 Largest Hotel Brands Estimated 6 Systemwide arketing Spend: $3.7bn Including Rewards Spend: $8.9bn EXPE / PCLN 6 arketing Spend: $7.3bn $6, believe it will likely have to compete in the SE channel for nonbranded keywords ("New York hotels," "Positano bed & breakfast," etc.), which could prove to have a steep learning curve and be an # of Booking.com Hotel Properties Company data, organ Stanley Research expensive endeavor against the more experienced OTAs, who already spend ~X as much on marketing as the top 5 largest hotel Second, the OTAs Are Large and Experienced Paid Search chains combined. Spenders/Traffic Acquirers The OTAs' aren't standing still on app downloads either. As shown This rising selection, combined with the OTAs' performance mar- below, while has the most app downloads in the US over the keting expertise and sheer dollar ad spend (~$9bn annually and past 3+ years, Expedia, Hotels.com and Booking.com are the next growing) will prove to be difficult barriers for to surmount as most commonly downloaded apps. If we look outside of the US, it continues to grow and attempt to take share of travelers' hotel Booking.com is still.5x larger than and note too that its spend and budgets. As in any online business, scale matters, and we pace of downloads outside of the US has been accelerating. Said would argue the OTAs' scale advantage (PCLN set to generate another way, outside of the US, Booking.com is extending its ~5X more room nights in 7 than ) and traffic acquisi- traffic lead over. tion and user conversion expertise will be difficult to displace. Exhibit 5 We expect the global OTAs combined to spend over $9bn per year on direct advertising... Over the last 3+ years, is the # most downloaded app in the U.S. among major hotels and OTAs... Cumulative U.S. App Downloads Q:4-3Q:7, Global OTA Direct Advertising Spend ($ millions) Exhibit 54,7, 9, Expedia 8, 7,35 Hotels.com 5,73 6, 6 Booking.com 4,848 KAYAK 9 3,64 4,,6 trivago, 7 Priceline 6 Source: Number of Downloads (in illions) e 8e Source: Sensor Tower, organ Stanley Research Company Data, organ Stanley Research ORGAN STANLEY RESEARCH 7

28 Exhibit 55 Exhibit 57:...but # outside the U.S....but it is falling behind Booking.com internationally Cumulative International App Downloads Q:4-3Q:7 Cumulative International App Downloads Q:4-3Q:7 9 Booking.com 84 Number of Downloads (in illions) 8 55 trivago 46 Hotels.com Expedia 8 KAYAK Number of Downloads (in illions) Source: Sensor Tower, organ Stanley Research Booking.com Source: Sensor Tower, organ Stanley Research Exhibit 56: app download growth is strong in the US... Cumulative U.S. App Downloads Q:4-3Q:7 Number of Downloads (in illions) Booking.com Source: Sensor Tower, organ Stanley Research 8 Expedia Hotels.com 3Q:7 Q:7 Q:7 4Q:6 3Q:6 Q:6 Q:6 4Q:5 3Q:5 Q:5 Q:5 4Q:4 3Q:4 Q:4 Q:4 Expedia Hotels.com 3Q:7 Q:7 Q:7 4Q:6 9 3Q:6 8 Q:6 7 Q:6 6 4Q:5 5 3Q:5 4 Q:5 3 Q:5 4Q:4 3Q:4 Q:4 Q:4 Priceline

29 Appendix Exhibit 58: 9% of survey respondents who used in the last months were satisfied with their experience (flat from last year) Experience with Using (Among Past onths Users) % 9% Not at all satisfied 3 3 Somewhat dissatisfied Neutral 9% customer satisfaction 9% customer satisfaction 9% customer satisfaction 5 Somewhat satisfied 5 % Very satisfied % % Source: AlphaWise, organ Stanley Research Exhibit 59 Exhibit 6...and US users have skewed even wealthier over the past year US users skew wealthier than non- users data % of all US Respondents 3 % % % 9% % % % of all US Respondents that DID use over the past months % 9% % % 9% % % % % % $-$ $-$4 $4-$6 $6-$75 $75+ $-$ Decline to answer $-$4 $6-$75 $75+ Decline to answer Annual Income Distribution (in $ thousands) Annual Income Distribution (in $ thousands) users over past months 7 $4-$6 Not users over past months 7 Source: AlphaWise, organ Stanley Research Source: AlphaWise, organ Stanley Research Exhibit 6 users skew younger than non- leisure and business travelers user 3 % Total % US % UK % 9% 3 % FR % % % 4 9% 9% % % % 9% 9% % % % of Respondents by Age Group % Business traveler % Number of companions in group, when booking on Leisure traveler 3 % Exhibit 6: 9% of survey respondents who used in the last months traveled in groups of 5 or fewer DE % 3 % % Source: AlphaWise, organ Stanley Research or older (traveled alone) ore than people Source: AlphaWise, organ Stanley Research ORGAN STANLEY RESEARCH 9

30 Exhibit 63: Travelers book far in advance through travel agents and at the last minute by calling hotels directly How long far in advance travelers book, by channel Exhibit 64: Survey data indicates that the regulatory burden on hosts may be lessening hosts who have alter behavior because of changes to laws or regulations % % % % 9% % % % % % % % % % % % % % % % Yes, because of changes I have listed or rented out my residence more often using % week or less Total -4 weeks Vacation Rental Site -3 months Hotel website 4-6 months Called or walked into hotel 7- months Travel agent ore than months Online travel agency Exhibit 65 German users skew wealthier than non- users... 6 data % 9% % % % % % Up to Decline to answer % of all Germany Respondents that DID use over the past months % of all Germany Respondents users over past months % of all UK Respondents 6 data % 9% 9% % Decline to answer Annual Income Distribution (in $ thousands) users over past months Source: AlphaWise, organ Stanley Research 3 % % % % 9% % % % % % Decline to answer Annual Income Distribution (in $ thousands) Exhibit 68...and the distribution of upper-income UK users has remained roughly unchanged over the past year % % Up to Not users over past months % of all UK Respondents that DID use over the past months % Source: AlphaWise, organ Stanley Research Exhibit 67 UK users skew wealthier than non- users... 6 Not users over past months Source: AlphaWise, organ Stanley Research Annual Income Distribution (in $ thousands) No, I have not changed how I list or rent out my residence on Exhibit 66...and the distribution of upper-income German users has remained roughly unchanged over the past year 3 3 % Yes, because of changes I have stopped listing or renting out my residence using Source: AlphaWise, organ Stanley Research Source: AlphaWise, organ Stanley Research Yes, because of changes I have listed or rented out my residence less often using 5 6 % 7 % % 9% % 9%% % Annual Income Distribution (in $ thousands) Source: AlphaWise, organ Stanley Research Decline to answer

31 Exhibit 69 French users skew wealthier than non- users... % of all France Respondents that DID use over the past months % of all France Respondents 6 data 4 4 Exhibit 7...and the distribution of upper-income French users has remained roughly unchanged over the past year 3 3 % % % % % % % % Up to Decline to answer % % % % % % 9% 9% 75+ Decline to answer % Up to Annual Income Distribution (in $ thousands) users over past months 5 4 Not users over past months Annual Income Distribution (in $ thousands) Source: AlphaWise, organ Stanley Research Source: AlphaWise, organ Stanley Research Exhibit 7: Our survey data shows 5% of 's leisure users have substituted away from traditional hotels... Accommodation Alternatives that Replaced (Among Leisure Users Past onths) +9bp +68bp -86bp +439bp +53bp -bp -369bp -6p +48bp 5% 5% % % % % % % 9% 9% % % % Traditional hotel Bed & Breakfast Friends/family Extended stay Vacation rental Other Vacation hotel site Rentals 5 6 Corporate apartment Other 7 I would not have taken this trip if not for Source: AlphaWise, organ Stanley Research Exhibit 7...and of 's business users have substituted away from traditional hotels Accommodation Alternatives that Replaced (Among Business Users Past onths) +bp -5bp +33bp +488bp -69bp +654bp -8bp +3bp -5bp 39% 39% 3 3 % % % % % % Other I would not have taken this trip if not for % Traditional hotel Bed & Breakfast Friends/family Extended stay Vacation rental Other Vacation hotel site Rentals 5 6 Corporate apartment 7 Source: Company data, organ Stanley Research ORGAN STANLEY RESEARCH 3

32 Exhibit 73...including 5 of leisure users (+557bp Y/Y) over the next months... Accommodation Alternatives that Likely to Replace N (Among Leisure Users Next onths) % 9% +557bp -49bp +8bp +95bp -39bp -6bp +78bp +48bp +48bp % % % % % 9% % % % Traditional hotel Bed & Breakfast Friends/family Extended stay Other Vacation Vacation rental hotel Rentals site 5 6 Corporate apartment Other 7 I would not have taken this trip if not for Source: AlphaWise, organ Stanley Research Exhibit 74...and 5% of business users (+9bp Y/Y) over the next months. Accommodation Alternatives that Likely to Replace N (Among Business Users Next onths) +9bp +bp +5bp +483bp -33bp -7bp +46bp -+bp +3bp 4 5% % % % % % % Traditional hotel Bed & Breakfast Friends/family Extended stay hotel 5 Source: AlphaWise, organ Stanley Research 3 Other Vacation Rentals 6 Vacation rental site 7 Corporate apartment Other I would not have taken this trip if not for

33 odel and Impact to Hotel RevPAR Exhibit 75: organ Stanley's Impact odel: Forecasted US / Europe RevPAR if Didn't Exist Traditional Hotel odel Before Impact Se Traditional Hotel Demand (Room Nights, m) % Growth Europe hotel rooms % Growth US hotel rooms % Growth Se Traditional Hotel Supply (Room Nights, m) % Growth 4,696 5, ,95.9% 6,83.% 6.3.9% ,44.% 7E,98 3.% 6.38.% 5.3.% 4,4. 8E,995 3.% 6.45.% 5.4.% 4,67. 9E 3,85 3.% % 4,34. E 3,77 3.% 6.68.% 5.46.% 4,49..% Implied Occupancy (ex impact for forecast years) % Growth % 67..% 68.%. 69.%.% 7.%.% 7.%. 7.. Expected Rate Growth at Estimated Occupancy Levels. 6.%.9%..%.9%. Se RevPAR Growth (ex impact for forecast years) % ,57 Source: AlphaWise, STR; KG; organ Stanley Research Exhibit 76: organ Stanley's Impact odel: Key Assumptions for Future Usage Key Assumptions for Future Usage Addressable User arket (m travellers who book online) Y/Y % Growth % 7E 6 8E 9E 4 E 8 User etrics Times Stayed at an by Users Per Year Bull Case Base Case Bear Case Room Nights per User per Booking Number of Users Per Stay % 6.% 6.% 4.% 4.% 4.%.%.%.% 5.% 5.% 5.% 33.% % % 3.% 7. 8 bp 8 bp 8 bp 8 bp 8 bp 8 bp 3 bp 3 bp 3 bp 8 bp 5 bp 5 bp 8 bp bp 75 bp 8 bp 5 bp bp Estimated User Penetration Bull Case Base Case Bear Case Y/Y Change in User Penetration Bull Case Base Case Bear Case Estimated no. of Users (mn) Bull Case Base Case Bear Case Estimated no. of Room Nights (mn) Bull Case Base Case Bear Case Estimated % of Traditional Hotel Room Nights Bull Case Base Case Bear Case % Cannibalizing Traditional Hotel Demand Bull Case Base Case Bear Case %.%.%... 4.% 4.% 4.% % % 5.9% 4. 49% 49% 49% 5% 5% 5% 5 5 5% % 5 4 Source: AlphaWise, STR; KG; organ Stanley Research ORGAN STANLEY RESEARCH 33

34 Exhibit 77: organ Stanley's Impact odel: Impact on Hotels Impact of on Traditional Hotel Demand Hotel Industry Occupancy Traditional model (pre-) Bull Case Change from traditional model Actual Change in Occupancy Base Case Cannibalized nights Change from traditional model Actual Change in Occupancy Bear Case Change from traditional model Actual Change in Occupancy E 8E 9E E % 68.% % % 69.% 68.9% -.%. 68.9% -.%. 68.9% -.%. 7.% 69.% -.%.% % -..% 7.% 69.% -.9%.% % -. -.% 7.9% % -.% -.% -.% % x.x.x.x 3. 3.% -. 3.% -. 3.% -. 3.%.% -.%.% -.% % % % Expected Impact on Rate Growth Bull Case Base Case Bear Case Rate vs. Occupancy Growth ultiplier RevPAR Growth Traditional model (pre-) Bull Case Change from traditional model Base Case Change from traditional model Bear Case Change from traditional model Source: AlphaWise, STR; KG; organ Stanley Research Exhibit 78: Lodging EBITDA and ultiple Changes Base Case EBITDA Old Lodging C-Corps HLT $,5 AR 3,65 LQ 347 H 85 STAY 65 CHH 36 8 New % Chg $,57 3, % % % % -% % C-Corps Avg Lodging REITs HST $,475 LHO 34 XHR 98 DRH 53 SHO 349 Old 9 New $,67 3, $,34 3, % $, % % % % % $, $, % Chg '9 New vs. '8 Old EBITDA Old % - % - $,97 3, % % % % -% -% % % -% % EBITDA Used for Bear Case New % Chg Old $,79 3, % % -% $,39 3, Bull Case New $,36 4,7 39, $, $, % % % -% - $, $, % Chg Bear Case Old New ultiple Used for Base Case Old New Bull Case Old New 9% % -% 9.8x 9.8x 8.7x 8.9x 7.x 8.5x.5x.6x 9.x.x 7.5x.8x 4.5x 4.3x.x.x 9.5x 3.6x 4.5x 4.3x.x.x 9.5x 3.6x 7.3x 7.x.4x 4.x.8x 7.6x 7.3x 7.3x.x 4.x.x 6.7x 8.8x 9.9x.5x.5x 4.8x 4.6x % % % - % 9.x 8.7x 8.8x 8.6x 9.x 9.9x.x 9.4x 9.9x.x.x.5x.3x.5x.5x.x.5x.3x.5x.5x 4.x 4.6x 3.8x 4.x 4.x 4.x 4.5x 3.x 3.8x 4.x REITs Avg % % % -% -% 8.9x 9.9x.8x.8x 4.x 3.9x Lodging Average % % % 8.8x 9.9x.x.x 4.5x 4.3x Source: Company data, organ Stanley Research 34

35 What Is? Founded in August 8 and based in San Francisco, California, is an online marketplace for people to list, discover, and book unique accommodations around the world. offers more than 3 million accommodations in more than 65, cities in over 9 countries. is privately owned, and according to press reports, was most recently valued at $3bn (The Wall Street Journal, 3/9/7). Exhibit 79: Screenshot of website interface Source:.com ORGAN STANLEY RESEARCH 35

36 How does make money? What are the key controversies around? charges two separate fees: a host service fee and a guest ser- Rental regulations: In some U.S. states and international regions, it vice fee. is illegal to rent out one's home/apartment on. This could be a limiting factor on 's supply growth. For reference, we wrote charges hosts a host service fee every time a booking is about this extensively in our report, Lodging and Internet: completed on 's online platform, based on the complete price Regulatory Guide for the Top 5 US Lodging arkets (4 Aug 7). of a reservation before any other fees have been added. Safety: is a marketplace that connects hosts and guests, and charges guests a -% guest service fee every time a as such, disclaims liability for the conduct of individual hosts and the booking is completed on 's online platform, based on the com- activities that occur at their properties. In contrast, hotels are subject plete price of a reservation before any other fees have been added. to stricter safety regulations (i.e., required to have smoke/fire The exact percentage is determined by the size of the reservation, alarms) and are liable for the well-being of their guests. and declines on a sliding scale from % to as the reservation cost increases. Consistency: is unique by design. Therefore, guests must rely on the description and pictures on the host's listing page for details Who competes with? about the property (size, amenities, cleanliness, noise level, etc.)...and they may not be exactly as advertised. Hotels offer more competes with most accommodation/lodging providers such predictability and certainty. as traditional hotels and motels, extended stay hotels, online travel agencies (OTAs), bed & breakfasts, vacation rentals, and other alter- Convenience: Renting on can be a less convenient travel native accommodation providers. experience than staying at a hotel, from the logistics of the host/ guest room key exchange to the absence of a concierge or front desk. What differentiates from competitors? Reliability: hosts can cancel a guest's reservation after it is Unique selection: One of 's main differentiating factors vs. confirmed. competitors is its unique and broad selection of accommodation properties. inventory ranges from the standard (entire apartments, entire homes, entire vacation rental property), to the extravagant (multi-million dollar mansion), to the most basic (a single room or a couch within a single room in an apartment). Authentic experience: A key value proposition is the opportunity to have a travel experience that is more unique and culturally immersive than staying at a chain hotel. For example, an user could visit a foreign country and stay in a local resident's home. Low prices: 's average price per room night (~$ in Europe, ~$6 in the US) is cheaper than the hotel average. For valuation methodology and risks associated with any price targets, ratings or recommendations referenced in this research report, please contact the Client Support Team as follows: US/Canada ; Hong Kong ; Latin America (U.S.); London +44 () ; Singapore ; Sydney +6 () ; Tokyo +8 () Alternatively you may contact your investment representative or organ Stanley Research at 585 Broadway, (Attention: Research anagement), New York, NY 36 USA. 36

37 ore on AlphaWise Source: AlphaWise Click here for more information about AlphaWise. ORGAN STANLEY RESEARCH 37