Mary Meeker's Annual Internet Trends Report

Thanks... KPCB Partners Especially Alex Tran / Dino Becirovic / Alexander Krey / Cindy Cheng who helped develop the ideas / presentation we hope you find useful... Hillhouse Capital Especially Liang Wu...his / their contribution of the China section of Internet Trends provides an especially thoughtful overview of the largest market of Internet users in the world... Participants in Evolution of Internet Connectivity From creators to consumers who keep us on our toes 24x7...and the people who directly help us prepare this presentation... Kara & Walt For continuing to do what you do so well... KPCB INTERNET TRENDS 2016 | PAGE 3

GLOBAL INTERNET TRENDS

Global Internet Users @ 3B Growth Flat = +9% vs. +9% Y/Y... +7% Y/Y (Excluding India) Source: United Nations / International Telecommunications Union, US Census Bureau. Internet user data is as of mid-year. Internet user data for: China from CNNIC, Iran from Islamic Republic News Agency, citing data released by the National Internet Development Center, India from IAMAI, Indonesia from APJII / eMarketer. KPCB INTERNET TRENDS 2016 | PAGE 5

Global Internet Users = 3B @ 42% Penetration... +9% vs. +9% Y/Y...+7% (Excluding India) Global Internet Users, 2008 – 2015 3,500 35% 3,000 30% ) M 2,500 25% M s ( h t ser w U 2,000 20% o et Gr n 1,500 15% % er /Y t Y n I al 1,000 10% b o l G 500 5% 0 0% 2008 2009 2010 2011 2012 2013 2014 2015 Global Internet Users Y/Y Growth (%) Source: United Nations / International Telecommunications Union, US Census Bureau. Internet user data is as of mid-year. Internet user data for: China from CNNIC, Iran from Islamic Republic News Agency, citing data released by the National Internet Development Center, India from IAMAI, Indonesia from APJII / eMarketer. KPCB INTERNET TRENDS 2016 | PAGE 6

India Internet User Growth Accelerating = +40% vs. +33% Y/Y... @ 277MM Users... India Passed USA to Become #2 Global User Market Behind China Source: United Nations / International Telecommunications Union, US Census Bureau. Internet user data is as of mid-year. Internet user data for: China from CNNIC, India from IAMAI. India users as of 10/2015 was 317MM per IAMAI; USA total population at 12/2015 (inclusive of all ages) was 323MM per US Census. KPCB INTERNET TRENDS 2016 | PAGE 7

India Internet Users = 277MM @ 22% Penetration... +40% vs. +33% Y/Y India Internet Users, 2008 – 2015 300 60% 250 50% ) M M 200 40% s ( h t w ser o U 150 30% Gr et n % er /Y t Y n I 100 20% a i d n I 50 10% 0 0% 2008 2009 2010 2011 2012 2013 2014 2015 India Internet Users Y/Y Growth (%) Source: IAMAI. Uses mid-year figures. KPCB INTERNET TRENDS 2016 | PAGE 8

Global Smartphone Users Slowing = +21% vs. +31% Y/Y Global Smartphone Unit Shipments Slowing Dramatically = +10% vs. +28% Y/Y Source: Nakono Research (2/16), Morgan Stanley Research (5/16). “Smartphone Users” represented by installed base. KPCB INTERNET TRENDS 2016 | PAGE 9

Global Smartphone User Growth Slowing... Largest Market (Asia-Pacific) = +23% vs. +35% Y/Y Smartphone Users, Global, 2005 – 2015 3,000 2015: Asia- Pacific = 52% 2,500 (MM) s r e 2,000 s U phone1,500 t r a m S 1,000 l oba 2008: Asia- l Pacific = 34% G 500 0 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 North America Western Europe Eastern Europe Asia-Pacific Latin America MEA Source: Nakono Research (2/16). * “Smartphone Users” represented by installed base. KPCB INTERNET TRENDS 2016 | PAGE 10

Global Smartphone Units Slowing Dramatically... After 5 Years of High Growth @ +10% vs. +28% Y/Y Smartphone Unit Shipments by Operating System, Global, 2007 – 2015 1,500 100% ) M M ( 1,200 80% s nt e pm ) hi 900 60% (% S th t w ni o U r G 600 40% Y / phone Y t r a m 300 20% S l oba l G 0 0% 2007 2008 2009 2010 2011 2012 2013 2014 2015 Android iOS Other Y/Y Growth Source: Morgan Stanley Research, 5/16. KPCB INTERNET TRENDS 2016 | PAGE 11

Android Smartphone Share Gains Continue vs. iOS... Android ASP Declines Continue...Delta to iOS @ ~3x Smartphone Unit Shipments, iOS vs. Android, Global, 2007 – 2016E 2015 Share: +7% Y/Y 1,200 iOS = 16% Android = 81% iOS Android (MM)800 ts n e m p i h S 400 t i n 2009 Share: U iOS = 14% -11% Y/Y Android = 4% 0 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016E iOS ASP ($) $594 $621 $623 $703 $712 $686 $669 $680 $717 $651 Y/Y Growth – 4% 0% 13% 1% -4% -2% 2% 5% -9% Android ASP – $403 $435 $441 $380 $318 $272 $237 $216 $208 Y/Y Growth – – 8% 1% -14% -16% -15% -13% -8% -4% Source: Morgan Stanley Research, 5/16. KPCB INTERNET TRENDS 2016 | PAGE 12

New Internet Users = Continue to be Harder to Garner Owing to High Penetration in Developed Markets KPCB INTERNET TRENDS 2016 | PAGE 13

With Already High Mobile Penetration in More Developed / Affluent Countries... New Users in Less Developed / Affluent Countries Harder to Garner, per McKinsey Countries fall into one of 5 groups based on Group 1: High barriers across the board; offline populations that are young, rural, and have low literacy barriers they face to Internet adoption Countries: Bangladesh, Ethiopia, Nigeria, Pakistan, Tanzania Offline population, 2014: 548 million Internet penetration, 2014: 18% Performance on Internet Group 1 Group 2: Medium to high barriers with larger challenges in incentives and Barriers Index Incentives Group 2 infrastructure; mixed demographics Average score 100 Group 3 Minimum - 0 Countries: Egypt, India, Indonesia, Philippines, Thailand Maximum -100 80 Group 4 Offline population, 2014: 1,438 million Group 5 Internet penetration, 2014: 20% 60 40 Group 3: Medium barriers with greatest challenge in incentives; rural and literate offline populations 20 Countries: China, Sri Lanka, Vietnam Offline population, 2014: 753 million Infrastructure 0 Internet penetration, 2014: 49% Group 4: Medium barriers with greatest challenge in low incomes and affordability; offline populations predominantly urban / literate / low income Low incomes Countries: Brazil, Colombia, Mexico, South Africa, Turkey and affordability Offline population, 2014: 244 million Internet penetration, 2014: 52% Group 5: Low barriers across the board; offline populations that are highly User capability literate and disproportionately low income and female Countries: Germany, Italy, Japan, Korea, Russia, USA 3 Offline population, 2014: 147 million Internet penetration, 2014: 82% Source: World Bank; McKinsey analysis from Internet Barriers Index KPCB INTERNET TRENDS 2016 | PAGE 14

Smartphone Cost in Many Developing Markets = Material % of Per Capita Income... 15% (Vietnam) / 10% (Nigeria) / 10% (India) / 6% (Indonesia), per McKinsey Average retail price of a smart phone, $USD, 2014 Ethiopia $262 47.6 Tanzania $198 21.5 Bangladesh $123 Cost of smartphone as a % 11.4 Vietnam $279 x% of GNI per capita, 2014 14.8 India $158 Developing Developed 10.1 Nigeria $307 10.3 Egypt $195 6.1 Indonesia $212 5.8 Philippines $163 4.7 Thailand $273 4.7 Colombia $291 3.7 South Africa $256 3.8 China $243 3.3 Turkey $522 4.8 Brazil $319 2.7 Russia $232 1.8 Mexico $244 2.5 Germany $486 1.0 Italy $327 0.9 Spain $269 0.9 South Korea $216 0.8 Japan $232 0.6 Source: McKinsey, Euromonitor, (smartphone prices); World Bank, estimates (GNI p.c., Atlas method) Note: Reflects true prices as paid by the consumer at point-of-sale; includes taxes and subsidies. Excludes data plan costs. KPCB INTERNET TRENDS 2016 | PAGE 15

GLOBAL MACRO TRENDS

Global Economic Growth = Slowing KPCB INTERNET TRENDS 2016 | PAGE 17

Global GDP Growth Slowing = Growth in 6 of Last 8 Years @ Below 20-Year Average Global Real GDP Growth (%), 1980 – 2015 6% 5% ) % 20-Year Avg ( h = 3.8% t 4% w o 35-Year Avg Gr = 3.5% P 3% l GD a e 2% l R a b Glo 1% 0% (1%) 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Source: IMF WEO, 4/16. Stephen Roach, “A World Turned Inside Out,” Yale Jackson Institute for Global Affairs, 5/16. Note: GDP growth based on constant prices (real GDP growth). KPCB INTERNET TRENDS 2016 | PAGE 18

Commodity Price Trends = In Part, Tell Tale of Slowing Global Growth KPCB INTERNET TRENDS 2016 | PAGE 19

Commodity Prices Down = -39% Since 5/14 vs. -8% Annual Average (5/11-4/14) & +6% (1/00-4/11) Global Commodity Prices, Bloomberg Commodity Index (Indexed to 0 @ 1/00), 2000 – 2016YTD 200% 150% x nde I 00) y t 1/ odi @ 100% m 0 om o t g C 50% r exed be d n I ( oom l B 0% (50%) 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 Source: Morgan Stanley, Bloomberg as of 5/25/16 Note: Bloomberg Commodity Index represents 22 globally traded commodities, weighted as: 31% Energy, 23% Grains, 17% Industrial Metals, 16% Precious Metals, 7% Softs (Sugar, Coffee, Cotton), and KPCB INTERNET TRENDS 2016 | PAGE 20 6% Livestock.

Global Growth Engines = Evolve Over Time KPCB INTERNET TRENDS 2016 | PAGE 21

Global Growth Engines @ ~2/3 of Global GDP Growth... 1985 = N. America + Europe + Japan 2015 = China + Emerging Asia Real GDP Growth Contribution by Region, 1985 / 2015 (Based on Purchasing Power Parity) 1985 2015 $19T = World GDP $114T = World GDP +4% Y/Y +3% Y/Y 0% N. America + 9% 15% Europe + Japan = 29% of Total China + 9% 9% 22% Emerging Asia = 26% 13% 18% of Total 7% N. America + 11% Europe + Japan = 28% 63% of Total 1% 13% China + 37% Emerging Asia = 63% of Total Europe N. America Japan China Emerging Asia (ex-China) Lat Am Middle East, Africa, Other Source: IMF WEO, 4/16. GDP growth based on constant prices (real GDP growth). PPP = Purchasing Power Parity exchange rate, national currency per international dollar. GDP PPP = GDP adjusted by PPP rate. Emerging Asia includes Bangladesh, Cambodia, India, Indonesia, Lao, Malaysia, Mongolia, Myanmar, Nepal, Philippines, Sri Lanka, Thailand, Vietnam and others and excludes China. KPCB INTERNET TRENDS 2016 | PAGE 22 GDP growth contribution based on annual snapshots stated above and not necessarily reflective of secular trends.

China’s Gross Capital Formation (Capital Equipment / Roads / Buildings...) Past 6 Years > Previous 30 Years KPCB INTERNET TRENDS 2016 | PAGE 23

China Gross Capital Formation = Slowing... Sum of Past 6 Years > Previous 30 Years China Gross Capital Formation, 1980 – 2015 (In 2010 Dollars) $4,500 $20T+ $21T+ ) $4,000 B $ ( $3,500 on i t a $3,000 m For $2,500 l a t pi $2,000 a C s $1,500 os r $1,000 G na hi $500 C $0 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 China Gross Capital Formation ($B) Source: China National Bureau of Statistics, 5/16. Assumes constant FX rate RMB/USD @ 6.5. Amounts are inflation adjusted to 2010 dollars based on IMF data on inflation rates (yearly average). KPCB INTERNET TRENDS 2016 | PAGE 24 Gross capital formation = gross fixed capital formation (majority) + changes in inventory. Gross fixed capital formation includes land improvements (fences, ditches, drains, and so on); plant, machinery, and equipment purchases; and the construction of roads, railways, and the like, including schools, offices, hospitals, private residential dwellings, and commercial and industrial buildings. It also includes the value of draught animals, breeding stock and animals for milk, for wool and for recreational purposes, and newly increased forest with economic value.

Shanghai Area Over Past 2+ Decades = Illustrates Magnitude of China (& Emerging Asia) Growth Shanghai, China, Pudong District 1987 2016 Source: Reuters/Stringer, Carlos Barria, Yichen Guo. KPCB INTERNET TRENDS 2016 | PAGE 25

Re-Imagination of China Over Past 3+ Decades – World’s Population Leader + #3 in Land Mass – Helped Drive Incremental Global Growth of Likes Which is Difficult to Repeat KPCB INTERNET TRENDS 2016 | PAGE 26

Interest Rates Have Fallen to Historically Low Levels = Interest Rate Trends = Can be Indicative of Perception for Growth Outlook KPCB INTERNET TRENDS 2016 | PAGE 27

USA 10-Year Treasury Yield = Low by Historical Standards USA 10-Year Treasury Yields, Nominal and Real, 1962 – 2016YTD 20% 15% ) % ( d 10% el i Y ear Y - 5% 10 0% (5%) 1962 1967 1972 1977 1982 1987 1992 1997 2002 2007 2012 Nominal Yield (%) Real Yield (%) Source: Morgan Stanley, Bloomberg, 5/16 Note: Real rates based on USGGT10Y Index on Bloomberg, which measures yield to maturity (pre-tax) on Generic 10-Year USA government inflation-index bonds. KPCB INTERNET TRENDS 2016 | PAGE 28

Global 10-Year Treasury Yields = Have Trended Down 10-Year Real Sovereign Bond Yields (%), Various Countries, 2001 – 2016YTD 8% ) % s ( 6% d el i Y d n o 4% B n g ei er v 2% o S eal R ear 0% Y - 10 (2%) 2001 2003 2005 2007 2009 2011 2013 2015 USA Canada UK Japan France Germany Italy Source: Morgan Stanley, Bloomberg, 5/16. Note: Real rates based on yield to maturity on 10-year inflation-indexed treasury security for each country. KPCB INTERNET TRENDS 2016 | PAGE 29

Total Global Debt Loads Over 2 Decades = High & Rising Faster Than GDP KPCB INTERNET TRENDS 2016 | PAGE 30

Global Government Debt @ 66% Average Debt / GDP (2015) & Up... +9% Annually Over 8 Years vs. +2% GDP Growth* for 50 Major Countries Global Debt By Type ($T, Constant 2014 FX), Q4:00 – Q2:15 Compound annual growth rate (%) $208 2000–2007 2007–Q2:15 +70T $41 Household 8.5 3.0 $138 $59 Corporate 5.7 6.4 $33 $84 $37 $19 $59 Government 5.9 8.7 $25 $32 $21 $37 $49 Financial 9.6 3.7 $20 Q4:00 Q4:07 Q2:15 Total debt as 250 274 299 GDP 4.1 2.2 % of GDP Growth*: Source: McKinsey Global Institute (3/16), IMF. *GDP growth rate based on constant prices and calculated as average of average growth rates across 50 countries from 2000-2007 and 2008-2015. KPCB INTERNET TRENDS 2016 | PAGE 31

Total Debt-to-GDP Ratios = High & Up in Most Major Countries... @ 202% Average vs. 147% (2000)* Change in Real Economy Debt / GDP (%), 2007 – Q2:15 140 Leveraging 130 Developed Emerging Increasing leverage Ireland 120 Singapore Hong Kong ) (% 90 China Greece Portugal P D 80 Belgium G France t / b 70 Finland e D Australia Canada Spain y 60 Slovakia m Colombia Malaysia Netherlands Japan o Italy n 50 Turkey Thailand Sweden o c Russia Brazil Chile Korea United Kingdom E 40 l a Mexico Switzerland Denmark e 30 Poland R Czech Republic Austria Norway n Morocco i Hungary e 20 Peru United States g Indonesia n South Africa Vietnam a 10 Philippines h Saudi Arabia Romania C Germany 0 Nigeria India -10 Argentina Egypt Deleveraging -20 Deleveraging 0 30 60 90 120 150 180 210 240 270 300 330 360 390 420 Q2:15 Real Economy Debt / GDP (%) Source: McKinsey Global Institute (3/16). Debt includes that owed by households, non-financial corporates, and governments (i.e. excludes financial sector debt). *Country inclusion per McKinsey; includes top developed countries by GDP and representative geographic selection of emerging countries. KPCB INTERNET TRENDS 2016 | PAGE 32

Demographic Trends = Slowing Population Growth... Slowing Birthrates + Rising Lifespans KPCB INTERNET TRENDS 2016 | PAGE 33

World Population Growth Rate Slowing = +1.2% vs. +2.0% (1975) Global Population and Y/Y % Growth, 1950 – 2050E 10 2.5% 8 2.0% ) B ) on ( (% i t 6 1.5% a te a opul R P th l 4 1.0% w o oba r l G G Y / Y 2 0.5% 0 0.0% Global Population (B) Y/Y Growth (%) Source: U.N. Population Division Note: Growth Rates based on CAGRs over 5 Year Periods. KPCB INTERNET TRENDS 2016 | PAGE 34

Global Birth Rates = Down 39% Since 1960 (1% Annual Average Decline) Birth Rates per 1,000 People per Year, By Region, 1960 – 2014 50 ear Y 40 er p e, l p eo 30 P 000 1, 20 er p e at R h t 10 r i B 0 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2014 World USA China India Europe / Central Asia East Asia / Pacific Middle East / North Africa Sub-Saharan Africa Source: World Bank World Development Indicators Note: Represents birth rates per 1,000 people per year. KPCB INTERNET TRENDS 2016 | PAGE 35

Global Life Expectancy @ 72 Years = Up 36% Since 1960 (0.6% Annual Average Increase) Life Expectancy (Years, Both Genders), By Region, 1960 – 2014 80 70 s) ear Y ( 60 cy an ect xp 50 e E f i L 40 30 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2014 World USA China India Europe / Central Asia East Asia / Pacific Middle East / North Africa Sub-Saharan Africa Source: World Bank World Development Indicators KPCB INTERNET TRENDS 2016 | PAGE 36

Net, Net, Economic Growth Slowing + Margins for Error Declining = Easy Growth Behind Us KPCB INTERNET TRENDS 2016 | PAGE 37

5 Epic Growth Drivers Over Past 2 Decades = Losing Mojo 1) Connectivity Growth Slowing – Internet Users rose to 3B from 35MM (1995) 2) Emerging Country Growth Slowing – Underdeveloped regions developed – including China / Emerging Asia / Middle East which rose to 69% of global GDP growth from 43%... 3) Government Debt Rising (& High) – Spending rose to help support growth...Government debt-to-GDP rose to 66% from 51% (2000) for 50 major economies 4) Interest Rates Have Declined – Helped fuel borrowing – USA 10-Year Nominal Treasury Yield fell to 1.9% (2016) from 6.6% (1995) 5) Population Growth Rate Slowing & Population Aging – Higher birth rates helped drive labor force growth – population growth rate continued to fall – to 1.2% from 1.6% (1995) Source: US Census, ITU, IMF, Stephen Roach, McKinsey, Bloomberg, US Bureau of Labor Statistics, UN Population Division KPCB INTERNET TRENDS 2016 | PAGE 38

Several Up / Down Cycles in Past 2 Decades = Internet 1.0 (2000)...Property / Credit (2008)... Stock / Commodity Markets Performance (% Change From 1/93), 1/93 – 5/16 800% 700% ) 600% 100% 500% 1993 =400% 1/ 1/ e ( 300% u al V 200% ex d n I 100% 0% 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 S&P 500 NASDAQ China Shanghai Composite MSCI Europe Source: Capital IQ. Note: All values are indexed to 1 (100%) on Jan 1, 1993. Data as of 5/2716. KPCB INTERNET TRENDS 2016 | PAGE 39

Adjusting to Slower Growth + Higher Debt + Aging Population Creates Rising Risks... Creates Opportunities for Businesses that Innovate / Increase Efficiency / Lower Prices / Create Jobs – Internet Can Be @ Core of This... KPCB INTERNET TRENDS 2016 | PAGE 40

ADVERTISING / COMMERCE + BRAND TRENDS

Online Advertising = Mobile + Majors + Newcomers Continue to Crank Away KPCB INTERNET TRENDS 2016 | PAGE 42

USA Internet Advertising Growth = Accelerating, +20% vs. +16% Y/Y... Owing to Mobile (+66%) vs. Desktop (+5%) USA Internet Advertising, 2009 – 2015 $70 35% $60 $60 30% ) B $ ( $50 ng $50 25% i $43 s i ) t r $37 e $40 20% (% dv $32 th A w t $30 15% o ne $26 r r G e $23 Y nt / I $20 10% Y A S U $10 5% $0 0% 2009 2010 2011 2012 2013 2014 2015 Desktop Advertising Mobile Advertising Y/Y Growth Source: 2015 IAB / PWC Internet Advertising Report. KPCB INTERNET TRENDS 2016 | PAGE 43

Google + Facebook = 76% (& Rising) Share of Internet Advertising Growth, USA Advertising Revenue and Growth Rates (%) of Google vs. Facebook vs. Other, USA, 2014 – 2015 35,000 $35 +18% Y/Y 30,000 $30 ) Others 25,000 $B $25 +13% Y/Y e ( u en 20,000 ev $20 R ng i s i 15,000 t $15 r e dv +59% Y/Y A 10,000 A $10 S U $5,000 $5 $0 $ 2014 2015 2014 2015 2014 2015 Google Facebook Others Source: IAB / PWC 2015 Advertising Report, Facebook, Morgan Stanley Research Note: Facebook revenue include Canada. Google USA ad revenue per Morgan Stanley estimates as company only discloses total ad revenue and total USA revenue. “Others” includes all other USA KPCB INTERNET TRENDS 2016 | PAGE 44 internet (mobile + desktop) advertising revenue ex-Google / Facebook.

@ Margin... Advertisers Remain Over-Indexed to Legacy Media % of Time Spent in Media vs. % of Advertising Spending, USA, 2015 Time Spent Ad Spend 50% e m i T on ng Total Of Which i 40% Internet Ad Mobile Ad pt ndi um pe 39% = $60B = $21B S 36% ons ng 30% C i a s i t di r e e 25% ~$22B M dv l 20% 22% 23% a A Opportunity ot or T in USA 16% of 10% 13% 12% % 10% 4% 0% Print Radio TV Internet Mobile Source: Advertising spend based on IAB data for full year 2015. Print includes newspaper and magazine. Internet includes desktop + laptop + other connected devices. ~$22B opportunity calculated assuming Mobile ad spend share equal its respective time spent share. Time spent share data based on eMarketer 4/16. Arrows denote Y/Y shift in percent share. KPCB INTERNET TRENDS 2016 | PAGE 45 Excludes out-of-home, video game, and cinema advertising.

Online Advertising Efficacy = Still Has Long Way to Go Google Has Proven Effective Online Advertising Works... Google = $75B Revenue (2015), +14% Y/Y / $510B Market Value (5/31/16) ...But Many Online (Video) Ads are Ineffective, per Unruly... 81% = Mute Video Ads 62% = Annoyed with / Put Off by Brand Forcing Pre-Roll Viewing 93% = Consider Using Ad Blocking Software ...But There are Ways Video Ads Can Work, per Unruly 1) Authentic 2) Entertaining 3) Evoke Emotion 4) Personal / Relatable 5) Useful 6) Viewer Control 7) Work with Sound Off 8) Non-Interruptive Ad Format Source: CapitalIQ as of 5/31/16, Unruly Future Video Survey, July 2015. N = 3,200 internet users surveyed from the US, UK, Germany, Australia, Sweden, France, Indonesia and Japan. KPCB INTERNET TRENDS 2016 | PAGE 46

Adblocking @ ~220MM Desktop Users (+16% Y/Y)...~420MM+ Mobile (+94%)... Majority in China / India / Indonesia = Call-to-Arms to Create Better Ads, per PageFair Global Adblocking Users on Web (Mobile + Desktop), 4/09 – 3/16 500 400 (MM) s ser 300 U ng i k oc 200 dbl A l a b 100 Glo 0 2009 2010 2011 2012 2013 2014 2015 Desktop Adblocking Software Users Mobile Adblocking Browser Users Source: PageFair, 5/16. Dotted line represents estimated data. These two data sets have not been de-duplicated. The number of desktop adblockers after 6/15 are estimates based on the observed trend in desktop adblocking and provided by PageFair. Note that mobile adblocking refers to web / browser-based adblocking and not in-app adblocking. Desktop adblocking estimates are for global monthly active users of desktop adblocking software between 4/09 – 6/15, as calculated in the PageFair & Adobe 2015 Adblocking Report. Mobile adblocking KPCB INTERNET TRENDS 2016 | PAGE 47 estimates are for global monthly active users of mobile browsers that block ads by default between 9/14 – 3/16, including the number of Digicel subscribers in the Caribbean (added 10/15), as calculated in the PageFair & Priori Data 2016 Adblocking Report.

Video Ads that Work = Authentic / Entertaining / In-Context / Often Brief Snapchat’s 3V Advertising Vertical (Made for Mobiles) / Video (Great Way to Tell Story) / Viewing (Always Full Screen) Spotify (10-Second Ad) in... Furious 7 (10-Second Ad) in... Snapchat Live Stories + Discover Ultra Music Festival Miami Live Story 26MM+ Views, 12/15 14MM+ Views, 3/15 +30% Lift in Subscription Intent, +3x Attendance Among Target Demo for 2x More Effective Than Snapchatters vs. Non-Snapchatters Typical Mobile Channels = Opening Weekend Box Office Source: Snapchat KPCB INTERNET TRENDS 2016 | PAGE 48

Commerce + Brands = Evolving Rapidly By / For This Generation KPCB INTERNET TRENDS 2016 | PAGE 49

Each Generation Has Slightly Different Core Values + Expectations... Shaped by Events that Occur in Their Lifetimes KPCB INTERNET TRENDS 2016 | PAGE 50

Consumer Preference / Value Evolution by Generation, USA... Millennials = More Global / Optimistic / Tolerant..., per Acosta Silent Baby Boomers Gen X Millennials Birth Years 1928 – 1945 1946 – 1964 1965 – 1980 1981 – 1996 Year Most of Generation 18-33 Years Old 1963 1980 1998 2014 Summary • Grew up during Great • Grew up during time of idealism • Grew up during time of change • Grew up during digital era with Depression with TV + car for every suburban politically, socially + economically internet, mobile computing, social • Fought 2nd “war to end all home • Experienced end of the Cold War, media + streaming media on wars” • Apollo, Civil Rights, Women’s Reaganomics, shift from iPhones • Went to college on G.I. Bill Liberation manufacturing to services • Experiencing time of rising • Raised “nuclear” families in time • Disillusionment set in with economy, + AIDS epidemic globalization, diversity in race + of great prosperity + Cold War assassination of JFK, Vietnam • Rise of cable TV + PCs lifestyle, 9/11, war on terror, mass War, Watergate + increase in murder in schools + the Great divorce rates Recession Core Values • Discipline • Anything is possible • Independent • Globally minded • Dedication • Equal opportunity • Pragmatic • Optimistic • Family focus • Question authority • Entrepreneurial • Tolerant • Patriotism • Personal gratification • Self reliance Work / Life Balance • Work hard for job security • Climb corporate ladder • Work / life balance important • Expanded view on work / life • Family time not first on list • Don’t want to repeat Boomer balance including time for parents’ workaholic lifestyles community service + self- development Technology • Have assimilated in order to • Use technology as needed for • Technology assimilated • Technology is integral keep in touch and stay informed work + increasingly to stay in seamlessly into day-to-day life • Early adopters who move touch through social media such technology forward as Facebook Financial Approach • Save, save, save • Buy now, pay later • Cautious, conservative • Earn to spend Source: Acosta Inc., Pew Research Image: Doomsteaddiner.net, Billboard.com, Metro.co.uk Note: Data from Acosta as of 7/13. Pew Research Center tabulations of the March Current Population Surveys (1963, 1980, 1998, and 2014). Pew Research defines each generation and may differ from KPCB INTERNET TRENDS 2016 | PAGE 51 other sources as there are varying opinions on what years each generation begin and end.

Characteristic Evolution by Generation @ Peak Adult Years (18-33), USA... Millennials = More Urban / Diverse / Single... Silent Baby Boomers Gen X Millennials Birth Years 1928 – 1945 1946 – 1964 1965 – 1980 1981 – 1996 Year Most of Generation 1963 1980 1998 2014 18-33 Years Old Location When Ages 18-33 64% 68% 83% 86% Metropolitan as % Total Diversity When Ages 18-33 84% 77% 66% 57% White as % Total Marital Status When Ages 18-33 64% 49% 38% 28% Married as % Total Education by Gender When Ages 18-33 12% Male / 7% Female 17% Male / 14% Female 18% Male / 20% Female 21% Male / 27% Female % with Bachelor’s Degree Employment Status by Gender 78% Male / 38% Female 78% Male / 60% Female 78% Male / 69% Female 68% Male / 63% Female When Ages 18-33 Employed as % Total* Median Household Income ** N/A $61,115 $64,469 $62,066 When Ages 18-33 Population of Generation 35MM 61MM 60MM 68MM When Ages 18-33 Source: Pew Research Image: Doomsteaddiner.net, Billboard.com, Metro.co.uk Note: *Only shows those that were civilian employed (i.e. excludes armed forces, unemployed civilians, and those not in labor force). **Median household income shown in 2015 dollars. Pew Research KPCB INTERNET TRENDS 2016 | PAGE 52 Center tabulations of the March Current Population Surveys (1963, 1980, 1998, and 2014). Pew Research defines each generation and may differ from other sources as there are varying opinions on what years each generation begin and end.

Marketing Channels Evolve With Time... Shaped by Evolution of Technology + Media KPCB INTERNET TRENDS 2016 | PAGE 53

Each New Marketing Channel = Grew Faster... Internet > TV > Radio Advertising Expenditure Ramp by Channel, First 20 Years, USA, 1926 – 2015 (In 2015 Dollars) $70 $60 Internet ) B Television $ ( s $50 Radio e ur t ndi $40 pe x E $30 ng i s i t r $20 e dv A $10 $0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Years Source: McCann Erickson (1926-1979); Morgan Stanley Research, Magna, RAB, OAAA, IAB, NAA, PIB (1980-2015) Note: Data adjusted for inflation and shown in 2015 U.S. dollars. Television consists of cable and broadcast television advertising. Radio consists of network, national spot, local spot, and streaming audio KPCB INTERNET TRENDS 2016 | PAGE 54 advertising. Internet consists of mobile and desktop advertising.

Retailing Channels Evolve With Time... Shaped by Evolution of Technology + Distribution KPCB INTERNET TRENDS 2016 | PAGE 55

Evolution of Commerce Over Past ~2 Centuries, USA = Stores  More Stores  Malls  E-Commerce Corner / General Stores Supermarkets Discount Chains Wholesale Clubs 1800s 1930s 1950-60s 1970-80s Illustrative Generational Overlap Millennials Generation X Baby Boomers Silent Generation Department Stores Shopping Malls Superstores E-Commerce Mid-1800s 1950s 1960-80s 1990s Source: McKinsey Image: Wikipedia.org, Barnumlanding.com, Cbsd.org, Dwell.com, Rediff.com, Freep.com, Corporate.walmart.com, Zdnet.com KPCB INTERNET TRENDS 2016 | PAGE 56 Note: Millennials defined as those born between 1980 and 2000. In 2015, they are ages 15-35. Gen X defined as those born between 1965 and 1979. In 2015, they are ages 36-50. Boomers defined as those born between 1946-1964. In 2015, they are ages 51-70. Silents defined as those born between 1925 – 1945. In 2015, they are ages 71 – 90. Note there are varying opinions on what years each generation begin and end.

New / Emerging Retailers Optimize for Generational Change = J.C. Penney  Meijer  Walmart  Costco  Amazon  Casper Retail Companies Founded by Decade (Illustrative Example), USA, 1900 – 2015 Generational Overlap Generation Z Millennials Generation X Baby Boomers Silent Generation GI Generation 1900s 1910s 1920s 1930s 1940s 1950s 1960s 1970s 1980s 1990s 2000s 2010s ` Source: KPCB, Retailindustry.about.com (1900s – 1980s), Ranker (1990s), Internet Retailer “2016 Top 500 Guide” (2000s – 2010s) Note: Companies shown above in chronological order by founding year by decade. Companies from 2000s onwards selected as diverse set of fast-growing companies based on web sales data from the Internet Retailer “2016 Top 500 Guide.” Gen Z defined as those born after 2000. In 2015, they are ages 0-15. Millennials defined as those born between 1980 and 2000. In 2015, they are ages 15-35. Gen X KPCB INTERNET TRENDS 2016 | PAGE 57 defined as those born between 1965 and 1979. In 2015, they are ages 36-50. Boomers defined as those born between 1946-1964. In 2015, they are ages 51-70. Silents defined as those born between 1925 – 1945. In 2015, they are ages 71 – 90. GI Generation defined as those born between 1900 – 1924. In 2015, they are age 91 – 115. Note there are varying opinions on what years each generation begin and end.

Millennials = Impacting + Evolving Retail... KPCB INTERNET TRENDS 2016 | PAGE 58

Millennials @ 27% of Population = Largest Generation, USA... Spending Power Should Rise Significantly in Next 10-20 Years Population by Age Range, USA, 2014 Household Expenditure, Annual Average, by Age of Reference Person, USA, 2014 70 $70 60 $60 Millennials ) K 50 $ $50 ( e ur (MM) 40 t $40 on ndi i t pe a 30 x $30 E opul l P 20 $20 A nnua S A U 10 $10 0 5 5 $0 5 24 34 44 54 64 74 5 34 44 54 64 74 7 <1 o o o o o o >7 <2 o o o o o > 5 t 5 t 5 t 5 t 5 t 5 t 5 t 5 t 5 t 5 t 5 t 1 2 3 4 5 6 2 3 4 5 6 Source: U.S. Census Bureau “2010-2014 American Community Survey 5-Year Estimates”, Bureau of Labor Statistics “Consumer Expenditure Survey 2014” Note: Millennials defined as persons born between 1980 – 2000. There are varying opinions on what years each generation begin and end. KPCB INTERNET TRENDS 2016 | PAGE 59

Internet Continues to Ramp as Retail Distribution Channel = 10% of Retail Sales vs. <2% in 2000 E-Commerce as % of Total Retail Sales, USA, 2000 – 2015 12% $340B+ of E-Commerce Spend ) % ( 10% es al S l ai 8% et R f o 6% ce as % er m 4% m o C - E 2% 0% 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Source: U.S. Census Bureau, Federal Reserve Bank of St. Louis (5/16) Note: E-commerce and retail sales data are seasonally adjusted. Retail sales exclude food services, retail trade from gasoline stations, and retail trade from automobiles and other motor vehicles. KPCB INTERNET TRENDS 2016 | PAGE 60

Retail = Technology + Media + Distribution Increasingly Intertwined KPCB INTERNET TRENDS 2016 | PAGE 61

Retail – The New Normal = Drive Transaction Volume  Collect / Use Data  Launch New Products / Private Labels... Amazon – Private Label Brand Launches, 2004 – 2015 Outdoor Home Electronic Fashion Brands Furniture Goods Accessories Franklin & Freeman, Franklin Strathwood Pinzon AmazonBasics Tailored, James & Erin, Lark & Ro, North Eleven, Scout + Ro, Society 2004 2008 2009 New York 2015 % Total Amazon Purchasers % Total Amazon Purchasers % Total Amazon Purchasers % Total Amazon Purchasers Which Purchased Home & Which Purchased Household Which Purchased Which Purchased: Garden Products: Products: Electronics (<$50) Products: 11% 10% 21% Men’s Apparel – 12% Women’s Clothing – 9% Source: Internet Retailer, Bizjournals.com, Cowen & Company Internet Retail Tracker Image: Amazon.com, Milled.com KPCB INTERNET TRENDS 2016 | PAGE 62 Note: Purchaser data based on Cowen & Company consumer tracking survey (n= ~2,500), as of 3/16. Data shown is percentage of Amazon purchasers who purchased items from a specific category.

...Products Become Brands...Brands Become Retailers... Retailers Become Products / Brands...Retailers Come Into Homes... Less differentiation between products / brands / retailers as single products evolve into brands + consumers shop directly from brands + retailers leverage insights to develop own vertically-integrated brands...New distribution models emerging enabling direct-to-consumer commerce in the home... Brands    Retailers New DTC Products Brands Retailers Products / Brands Distribution Models (Casper) (Warby Parker) (Thrive Market) (Stitch Fix) Image: Myjane.ru, CNBC.com, Vanityfair.com, Insidebusinessnyc.com, Funandfit.org, Thrivemarket.com, Thedustyrosestyle.com, Stitchfix.tumblr.com KPCB INTERNET TRENDS 2016 | PAGE 63

...Physical Retailers Become Digital Retailers... Digital Retailers Become Data-Optimized Physical Retailers... Physical Retailers Evolving & Increasing E-Commerce Presence...New Products / Brands / Retailers Launching Physical Stores / Showrooms / Retail Channels...Omni-Channel is Key...Warby Parker @ $3K Annual Sales per Square Foot = One of Top Grossing Physical Retailers per Square Foot in USA Offline  Online Online  Offline (Neiman Marcus) (Warby Parker) 26% of F2015 Sales on Internet, +24% Y/Y 31 locations (5/16), up from 10 locations (12/14) Top 5 Physical Retailers by Sales / Sq. Ft., USA, 2015* Apple $5,546 Warby Parker $3,000 Tiffany & Co. $2,951 Lululemon $1,560 Athletica Michael Kors $1,466 Source: Company filings, Fast Company, Time, eMarketer Image: Pursuitist.com, Digiday.com, Warbyparker.com KPCB INTERNET TRENDS 2016 | PAGE 64 Note: *Excludes gas stations. Based on sales figures from trailing 12 months. Warby Parker figures as of February 2015.

...Connected Product Users Easily Notified When to Buy / Upgrade... Can Benefit from Viral Sharing Ring Connected Devices with Sharing of Events Captured on Ring Proliferation of Ring Connected Devices Sharable Content on Neighborhood Level – Nextdoor, TV... Serving Broader Communities April 2015 December 2015 May 2016 Source: Ring, Nextdoor, WLKY News Image: Ring.com, Whas11.com KPCB INTERNET TRENDS 2016 | PAGE 65

Internet-Enabled Retailers / Products / Brands On Rise = Bolstered by Always-On Connectivity + Hyper-Targeted Marketing + Images + Personalization KPCB INTERNET TRENDS 2016 | PAGE 66

Hyper-Targeted Marketing = Driving Growth for Retailers / Products / Brands Internet = Driving Force for New Product Introductions with Hypertargeting / Intent-Based Marketing via Facebook / Twitter / Instagram / Google... Stance Combatant Gentlemen ‘Our customer acquisition strategy was Facebook. Our [target customer] After noticing that its Instagram placements were outperforming all other typically spends a lot of time on Facebook...Every $100 we spent on placement types in its Star Wars collection launch campaign, Stance Facebook was worth $1,000 in sales. For us, it was a simple math decided to create a dedicated ad set to maximize its ad spend against this equation.’ placemen & build upon Instagram’s unique visual nature and strong targeting capabilities. ‘We target based on [Facebook] likes...For example, we have a lot of guys in real estate who are climbing up the ladder. What we do is we Stance targeted the ads to adults whose interests include the put these guys into cohorts and we say, ‘These are our real estate Star Wars movies, but excluded those interested only in specific guys.’ Star Wars characters. The ‘Sock Wars’ campaign generated an impressive 36% boost to return on ad spend. - Vishaal Melwani CEO and Founder, Combatant Gentlemen Source: One Million by One Million Blog, Instagram Business Image: Pinterest.com, Instagram Business KPCB INTERNET TRENDS 2016 | PAGE 67

Stitch Fix User Experience = Micro Data-Driven Engagement & Satisfaction... Data Collection + Personalization / Curation + Feedback... Stitch Fix = Applying Netflix / Spotify-Like Content Discovery to Fashion... Each Customer = Differentiated Experience...99.99% of Fixes Shipped = Unique Data-Driven Onboarding Process Ship ‘Fixes’ with Curated Items Customer Preferences & = Mix of Art + Science Based on Preferences / Style Feedback Collect data points on customer preferences / Allows clients to try products selected by stylists Collect information on customer experience to style / activities. 46% of active clients provide in comfort of home / return items they don’t like drive future product selection Pinterest profiles. Stylists use Pinterest boards + access to algorithms to help improve product selection Source: Stitch Fix Image: Forbes.com KPCB INTERNET TRENDS 2016 | PAGE 68

...Stitch Fix Back-End Experience = 39% of Clients Purchase Majority of Clothing from Stitch Fix vs. ~30% of Clients Y/Y Stitch Fix = Data On Users + Data on Items + Constantly Improving Algorithms = Drive High Customer Satisfaction...100% of Purchases from Recommendations Data Collection on Data Networking Effect... Strong Consumer Engagement / Item-by-Item Basis Coupled with Helps Stylist Predict Success of Anticipation...Increased Wallet User Insights Items with Specific Client Share... Stitch Fix captures 50-150 attributes on each The more information collected, the better the 39% of Stitch Fix clients get majority of item, uses algorithms + feedback to determine probability of success. Stitch Fix showing 1:1 clothing from Stitch Fix, up probability of success (i.e. item will be correlation between probability of purchase per from ~30% of clients a year ago purchased) for specific demographics, allows item and observed purchase rate over time stylists to better select items for clients Example of Product Success 100% Probability by Age d e 80% s a h c r u 60% P n o ti r o 40% Example of Product Success p o Probability by Sizing r P l a 20% tu c A 0% 0% 20% 40% 60% 80%100% Probability of Purchase Source: Stitch Fix Image: Cheapmamachick.com KPCB INTERNET TRENDS 2016 | PAGE 69

Many Internet Retailers / Brands @ $100MM in Annual Sales* in <5 Years... Took Nike = 14 Years / Lululemon = 9 / Under Armour = 8** Viral Marketing / Sharing Mechanisms (Facebook / Instagram / Snapchat / Twitter...) + On-Demand Purchasing Options via Mobile / Web + Access to Growth Capital + Millennial Appeal = Enabling Rapid Growth for New Products / Brands / Retailers Sales Growth For Select Internet Retailers*, USA, First 5 Years Since Inception $100 $80 ) M $M $60 es ( al S al $40 u n n A $20 Average $0 0 1 2 3 4 5 Year Since Inception Source: Internet Retailer “2016 Top 500 Guide”, company filings Note: *Data only for e-commerce sales and shown in 2015 dollars. **Years to reach $100MM in annual revenue in 2015 dollars. Chart includes pure-play e-commerce retailers and evolved pure-play retailers. Companies shown include Birchbox, Blue Apron, Bonobos, Boxed, Casper, Dollar Shave Club, Everlane, FitBit, GoPro, Harry’s, Honest Company, Ipsy, Nasty Gal, Rent the Runway, KPCB INTERNET TRENDS 2016 | PAGE 70 TheRealReal, Touch of Modern, and Warby Parker. The Top 500 Guide uses a combination of internal research staff and well-known e-commerce market measurement firms such as Compete, Compuware APM, comScore, ForeSee, Experian Marketing Services, StellaService and ROI Revolution to collect and verify information.

RE-IMAGINING COMMUNICATION VIA SOCIAL PLATFORMS – – VIDEO – IMAGE – MESSAGING

Visual (Video + Image) Usage Continues to Rise KPCB INTERNET TRENDS 2016 | PAGE 72

Millennial Social Network Engagement Leaders = Visual... Facebook / Snapchat / Instagram... Age 18-34 Digital Audience Penetration vs. Engagement of Leading Social Networks, USA, 12/15 1,200 or t i s i 1,000 V r pe s e 800 nut i M 600 y hl ont Snapchat M 400 Instagram ge a r 200 e v Twitter A 0 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% % Reach Among Age 18-34 Source: ComScore Media Metrix Multi-Platform, 12/15. KPCB INTERNET TRENDS 2016 | PAGE 73

Generation Z (Ages 1-20) = Communicates with Images Attributes – Millennials vs. Gen Z Millennials vs Gen Z Tech Savvy: 2 screens at once Tech Innate: 5 screens at once Communicate with text Communicate with images Curators and Sharers Creators and Collaborators Now-focused Future-focused Optimists Realists Want to be discovered Want to work for success Source: “Engaging and Cultivating Millennials and Gen Z,” Denison University and Ologie, 12/14. Note: Gen Z defined in this report as those born after 1995. In 2016 they are ages 1-20. Note that there may be different opinions on which years each generation begins and ends. KPCB INTERNET TRENDS 2016 | PAGE 74

Video Viewing Evolution Over Past Century = Live  On-Demand  Semi-Live  Real-Live KPCB INTERNET TRENDS 2016 | PAGE 75

Video Evolution = Accelerating Live (Linear)  On-Demand  Semi-Live  Real-Live Live (Linear) On-Demand Semi-Live Real-Live Traditional TV DVR / Streaming Snapchat Stories Periscope + Facebook Live 1926 1999 2013 2015 / 2016 Tune-In or Watch on Tune-In Within 24 Tune-In / Watch Miss Out Own Terms Hours or Miss Out on Own Terms Mass Concurrent Mass Disparate Mostly Personal Mass Audience, Audience Audience Audience yet Personal Real-Time Buzz Anytime Buzz Anytime Buzz Real Time + Anytime Buzz Images: Facebook, Twitter, Snapchat, Netflix, TiVopedia, BT.com 1926 - First television introduced by John Baird to members of the Royal Institution. 1999 - First DVR released by Tivo. 2013 – Snapchat Stories launched. KPCB INTERNET TRENDS 2016 | PAGE 76

Video Usage / Sophistication / Relevance Continues to Grow Rapidly KPCB INTERNET TRENDS 2016 | PAGE 77

User-Shared Video Views on Snapchat & Facebook = Growing Fast Facebook Daily Video Views, Snapchat Daily Video Views, Global, Q3:14 – Q3:15 Global, Q4:14 – Q1:16 10 10 ) ) B 8 B 8 ( ( ay D ay er D p er s 6 p 6 s ew i ew V i eo V d 4 eo 4 i d V i V 2 2 0 0 Q3:14 Q4:14 Q1:15 Q2:15* Q3:15 Q4:14 Q1:15 Q2:15 Q3:15 Q4:15 Q1:16 Source: Facebook, Snapchat. Q2:15 Facebook video views data based on KPCB estimate. Facebook video views represent any video shown onscreen for >3 seconds (including autoplay). Snapchat video views counted instantaneously on load. KPCB INTERNET TRENDS 2016 | PAGE 78

Smartphone Usage Increasingly = Camera + Storytelling + Creativity + Messaging / Sharing KPCB INTERNET TRENDS 2016 | PAGE 79

Snapchat Trifecta = Communications + Video + Platform... Stories (Personal)  Live (Personal + Pro Curation)  Discover (Pro) Stories (Personal) Live (Personal + Pro Curation) Discover (Professional) 10/13 Launch 6/14 1/15 Alex Alexander Dino   Cindy Anjney Arielle Aviv 10–20MM Snapchatters View 70MM+ Snapchatters View Live Stories Each Day Discover Each Month More Users Watched College Top Performing Channels Average Football and MTV Music Awards on 6 – 7 minutes per Snapchatter per Snapchat than Watched the Events Day on TV Source: Snapchat KPCB INTERNET TRENDS 2016 | PAGE 80

Advertisers / Brands = Finding Ways Into... Camera-Based Storytelling + Creativity + Messaging / Sharing KPCB INTERNET TRENDS 2016 | PAGE 81

Brand Filters Integrated into Snapchat Snaps by Users... Often Geo-Fenced, in Venue ‘Love at First Bite’ ‘World AIDS Day – Join the Fight’ by KFC by (RED) 9MM+ Views 76MM+ Views Geofilter offered @ 900+ KFCs Each time a geofilter was sent, Bill & Melinda Gates in UK and applied 200K+ times, Foundation donated $3 to (RED)’s fight against AIDS 12/15 – 2/16 12/15 +23% Visitation Lift Within 7 Days +90% Higher Likelihood of Donating to (RED) of Seeing Friend’s Geofilter Among Those Who Saw Geofilter Source: Snapchat KPCB INTERNET TRENDS 2016 | PAGE 82

Branded Snapchat Lenses & Facebook Filters... Increasingly Applied by Users Taco Bell Cinco de Mayo Lens Gatorade Super Bowl Lens Iron Man Filter from MSQRD 224MM Views on Snapchat 165MM Views on Snapchat 8MM+ Views on Facebook 5/5/16 2/7/16 3/9/16 Average Snapchatter Plays With Sponsored Lens for 20 Seconds Source: Snapchat, Facebook Time on sponsored lens excludes time taking and uploading image / video. KPCB INTERNET TRENDS 2016 | PAGE 83

Real-Live = Facebook Live... New Paradigm for Live Broadcasting KPCB INTERNET TRENDS 2016 | PAGE 84

UGC (User Generated Content) @ New Orders of Viewing Magnitude... Facebook Live = Raw / Authentic / Accessible for Creators & Consumers Candace Payne in Chewbacca Mask on Facebook Live Most Viewed Live Video @ 153MM+ Views, 5/16 Kohl’s = Mentioned 2 Times in Video Kohl’s = Became Leading App in USA iOS App Store Chewbacca Mask Demand Rose Dramatically Source: Facebook KPCB INTERNET TRENDS 2016 | PAGE 85

Live Sports Viewing = Has Always Been Social But.... It’s Just Getting Started KPCB INTERNET TRENDS 2016 | PAGE 86

How Often are You Able to Watch a Game (on Sidelines or TV) with All Your Friends Who Share Your Team Love? Live Streaming – Wrapped with Social Media Tools – Helps Make that More of a Reality... KPCB INTERNET TRENDS 2016 | PAGE 87

2016E = Milestone Year for ‘Traditional’ Live Streaming on Social Networks... NFL Live Broadcast TV of Thursday Night Football on Twitter (Fall 2016) Hypothetical Mock-Up Complete Sports Viewing Platform = Live Broadcast + Analysis + Scores + Replays + Notifications + Social Media Tools Vertical View = Horizontal View = Tune-In Notifications, Scoreboard Allows Fans to Live Broadcast + Tweets Unencumbered, Full- Game Reminders, Follow Play-by-Play Dashboard for Social Screen, TV-like Viewing Breaking Actions Media Engagement Experience Tweets Engage Fans in Professional Toggle Between Tweets Real-Time Conversation Commentary and from Stadium / Nearby / All Analysis Source: KPCB Hypothetical Mock-Up. Design provided by Brian Tran (KPCB Edge) KPCB INTERNET TRENDS 2016 | PAGE 88

Image Usage / Sophistication / Relevance Continues to Grow Rapidly KPCB INTERNET TRENDS 2016 | PAGE 89

Image Growth Remains Strong Daily Number of Photos Shared on Select Platforms, Global, 2005 – 2015 3,500 ) 3,000 M M ( Snapchat y 2,500 a Facebook Messenger (2015 only) D r Instagram d pe 2,000 WhatsApp (2013 onward only) e r Facebook ha 1,500 S os Facebook- hot 1,000 owned P properties of # 500 0 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Source: Snapchat, Company disclosed information, KPCB estimates Note: Snapchat data includes images and video. Snapchat stories are a compilation of images and video. WhatsApp data estimated based on average of photos shared disclosed in Q1:15 and Q1:16. Instagram data per Instagram press release. Messenger data per Facebook (~9.5B photos per month). Facebook shares ~2B photos per day across Facebook, Instagram, Messenger, and WhatsApp KPCB INTERNET TRENDS 2016 | PAGE 90 (2015).

Images = Monetization Options Rising KPCB INTERNET TRENDS 2016 | PAGE 91

Image-Based Platforms Like Pinterest = Often Used for Finding Products / Shopping... % of Users on Each Platform Who Utilize to ‘What Do You Use Pinterest For?’ Find / Shop for Products, USA, 4/16 (% of Respondents), USA, 4/16 60% 55% Viewing photos 60% 50% Finding / shopping for 55% products 40% Sharing photos / videos/ 24% personal messages 30% Watching videos 15% 20% 12% 12% News 10% 10% 9% 5% 3% 0% Networking / promotion 10% Pinterest Facebook Instagram Twitter LinkedIn Snapchat Source: Cowen & Company ”ShopTalk Conference Takeaways: A Glimpse Into The Future of Retail & eCommerce” (05/16) Note: Based on Cowen & Company proprietary Consumer Internet Survey from April / May 2016 of 2,500 US consumers, 30% of which where Pinterest MAUs as of 4/16. KPCB INTERNET TRENDS 2016 | PAGE 92

...Image-Based Platforms Like OfferUp = High (& Rising) Engagement Levels & Used for Commerce... Average Daily Time Spent per User, USA, 11/14 & 6/15 50 42 11/14 41 6/15 40 ay 30 D 25 25 25 er p 21 21 21 es 21 t 20 u 17 17 n i M 13 10 0 Facebook Instagram OfferUp Snapchat Pinterest Twitter Source: OfferUp, Cowen & Company “Twitter/Social User Survey 2.0: What’s changed?” Note: Based on SurveyMonkey survey conducted in June 2015 on 2,000 US persons aged 18+ KPCB INTERNET TRENDS 2016 | PAGE 93

...Image-Based Peer-to-Peer (P2P) Marketplace OfferUp = Ramping Faster than eBay @ Same Stage... OfferUp vs. eBay GMV Growth, First 8 Years Since Inception $16 $12 eBay OfferUp ) B $8 ($ MV G $4 $0 0 1 2 3 4 5 6 7 8 Year Since Inception Source: OfferUp, company filings, and KPCB estimates. Note: Shown on a calendar year basis and in nominal dollars. eBay was launched in 1995 and OfferUp in 2011. KPCB INTERNET TRENDS 2016 | PAGE 94

...Image-Based Platform Houzz = Content + Community + Commerce Continue to Ramp... Houzz – Content (Photos) / Community (Professionals + Consumers) / Commerce (Products), 4/12 – 4/16 40MM Consumers 5.5M Products Available for Purchase on Houzz Content (Photos) Houzz Marketplace 13MM Commerce (All Launched 10/14 Products) 10MM 5.5M Active Professionals 120K 1.2MM 400K 70K Source: Houzz 5.5MM products are available on Houzz for purchase directly within the app and on Houzz.com (Houzz Marketplace). KPCB INTERNET TRENDS 2016 | PAGE 95 There are 13MM total products available on Houzz Marketplace + linked to merchant sites.

...Houzz Personalized Planning with Images = 3-4x Higher Engagement...5x Higher Purchase Conversion View In My Room (2/16 Launch) Sketch (12/15) Pick a Product & Preview What It Looks Like Add Products from Houzz Marketplace to In Any Room Through Camera Any Photo on Houzz or Your Own Sketch 50% of Users Who Made a Purchase in Latest Version of Houzz App (Since 2/17/16) Over 500K Sketches Saved Since Launch Used View In My Room Users = 97% More Likely to Use Houzz Next Time Sketch Users = 5x More Likely to Purchase... They Shop...5.5x More Likely to Purchase... Spend 4x More Time in App Spend 3x More Time in App Source: Houzz KPCB INTERNET TRENDS 2016 | PAGE 96

Messaging = Evolving Rapidly KPCB INTERNET TRENDS 2016 | PAGE 97

Messaging Leaders = Strong User (+ Use) Growth KPCB INTERNET TRENDS 2016 | PAGE 98

Messaging Continues to Grow Rapidly... Leaders = WhatsApp / Facebook Messenger / WeChat Monthly Active Users on Select Social Networks and Messengers, Global, 2011 – 2015 1,200 WhatsApp ) 1,000 Launched 2010 M M ( Facebook Messenger s (2011) er 800 s U WeChat e (2011) v i t 600 c A Instagram y (2010) l h t 400 n Twitter o (2006) M 200 LinkedIn (2003) 0 2011 2012 2013 2014 2015 LinkedIn Twitter Instagram WhatsApp WeChat Facebook Messenger Source: Facebook, WhatsApp, Tencent, Instagram, Twitter, LinkedIn, Morgan Stanley Research Note: 2013 data for Instagram and Facebook Messenger are approximated from statements made in early 2014. Twitter users excludes SMS fast followers. KPCB INTERNET TRENDS 2016 | PAGE 99

Messaging = Evolving from Simple Social Conversations to More Expressive Communication... KPCB INTERNET TRENDS 2016 | PAGE 100

Messaging Platform Evolution = More Tools for Simple Self-Expression Global Electronic Messaging Platforms – Evolution of Simple Self-Expression Japanese Cell Phones – Type- AOL Instant Messenger – NTT DoCoMo- Apple iOS 5 – Based Emoji Convert Text Emoticon to Emoji Native Emoji 1990s Graphical Smiley 1999 2011 1997 Line – Bitstrips – Bitmoji Facebook Messenger – Snapchat – Stickers Personalized Emoji GIF Keyboard Lenses 2011 2014 2015 2015 Source: Wired, Company Statements, Press Releases. KPCB INTERNET TRENDS 2016 | PAGE 101

...Messaging = Evolving from Simple Social Conversations to Business-Related Conversations KPCB INTERNET TRENDS 2016 | PAGE 102

Asia-Based Messaging Leaders = Continue to Expand Uses / Services Beyond Social Messaging Name KakaoTalk WeChat LINE Launch March 2010 January 2011 June 2011 Primary Country Korea China Japan Banking / Financial Services Kakao Bank (11/15) WeBank (1/15) Debit Card (2016) Enterprise  Enterprise WeChat (3/16)  Online-To-Offline (O2O) Kakao Hairshop (1H:16E)  Grocery Delivery (2015) New Services Kakao Driver (1H:16E) Added 2015 -16* TV Kakao TV (6/15)  Line Live & Line TV (2015) Video Calls / Chat (6/15)   Taxi Services Kakao Taxi   (3/15) Messaging    Group Messaging    Voice Calls Free VoIP calls (2012) WeChat Phonebook  (2014) Previous Existing Payments KakaoPay (2014) (2013) Line Pay (2014) Services Stickers (2012) Sticker shop (2011) (2013) Games Game Center (2012) (2014) (2011) Commerce Kakao Page (2013) Delivery support Line Mall (2013) w / Yixun (2013) Media Kakao Topic (2014)   QR Codes  QR code  identity (2012) User Stories / Moments Kakao Story (2012) WeChat Moments Line Home (2012) Developer Platform KakaoDevelopers WeChat API Line Partner (2012) Source: Company websites, press releases, Morgan Stanley Research. *Blue shading denotes that at least one of the platforms listed has added new features since 2015. Some features for other platforms may have been added in prior years KPCB INTERNET TRENDS 2016 | PAGE Note: Enterprise denotes product made specifically for messaging or social networking within the enterprise, which is distinct from B2C messaging where businesses engage with current or potential 103 customers.

Messaging Secret Sauce = Magic of the Thread = Conversational... Remembers Identity / Time / Specifics / Preferences / Context Hyatt Rogers Communications Check Availability / Reservations / Order Room Service Ask Questions / Update Account / Set Up New Plan Started Offering Customer Service on Started Offering Customer Service on Facebook Messenger in 11/15 Facebook Messenger in 12/15 +20x Increase in Messages Received 65% Increase in Customer Satisfaction by Hyatt Within ~1 Month 65% Decrease in Customer Complaints Source: “Digital Transformation for Telecom Operators,” by Deloitte, 2016. Wired. KPCB INTERNET TRENDS 2016 | PAGE The Commissioner for Complaints for Telecommunications Services (CCTS) reported a 65 per cent decrease in customer complaints between 8/15 and 1/16 compared to the previous six months 104

Messaging Platforms = Millions of Business Accounts Helping Facilitate Customer Service & Commerce... Business / B2C Chat for Advertising Partnerships / Official Engagement Payments SMEs (Within Other Accounts Messengers) Services 10MM+ ~80% WeChat Pay Official Official Weidian Official Accounts Users Follow Official (2013) Accounts Accounts (2014) Accounts (2012) (2012) 1B+ Facebook Messages / Month Messaging via Shopify & 50MM+ Between Businesses and Payments Pages (2011) Sponsored Zendesk Small Business Users, +2x Y/Y (2015) Messages Partnership Pages 80% Chatbots (2016) (2015 / 2016) Businesses Active on Platform (2016) Mobile Official Accounts & 2MM+ Line Pay Line @ Official Commerce / Line@ + Official -- (2014) (2012 / 2015) Accounts Stores on Accounts (2012) Line@ (2016) Chatbots Platform (2016) Source: WeChat, Line, Facebook Messenger, various press releases, “WeChat’s Impact: A Report on WeChat Platform Data,” by Grata (2/15) KPCB INTERNET TRENDS 2016 | PAGE 105

...Messaging Platforms = Conversational Commerce Ramping Shopper in Thailand on Instagram Browsing Begins on Instagram...Conversation / Payment / Confirmation Ends on Line Visit Instagram Browse Inquire Get Confirm Ship & Shop Products About Payment Payment Track Order Product via Details Line Source: Commerce + Mobile: Evolution of New Business Models in SEA, 7/15. KPCB INTERNET TRENDS 2016 | PAGE 106

Best Ways for Businesses to Contact Millennials = Social Media & Chat... Worst Way = Telephone Popularity of Business Contact Channels, by Age Which channels are most popular with your age-profiled customers? (% of contact centers) % of Centers Reporting Most Popular Contact Channels by Generation Internet / Electronic Smartphone Web Chat Social Media Messaging Application Telephone (e.g. email, SMS) Generation Y 24% 24% 21% 19% 12% st st rd th th (born 1981-1999) (1 choice) (1 choice) (3 choice) (4 choice) (5 choice) Generation X 21% 12% 28% 11% 29% rd th nd th st (born 1961-1980) (3 choice) (4 choice) (2 choice) (5 choice) (1 choice) Baby Boomers 7% 2% 24% 3% 64% rd th nd th st (born 1945-1960) (3 choice) (5 choice) (2 choice) (4 choice) (1 choice) Silent 2% 1% 6% 1% 90% Generation rd th nd th st (3 choice) (4 choice) (2 choice) (5 choice) (1 choice) (born before 1944) Source: “Global Contact Center Benchmarking Report,” Dimension Data, 2015. N = 717 Contact Centers, Global. Results are shown based on contact centers that actually tracked channel popularity. Percentage may not add up to 100 owing to rounding. KPCB INTERNET TRENDS 2016 | PAGE Generation Y is typically referred to as “Millennials” 107

Android / iOS Home Screens (Like Portals in Internet 1.0) = Mobile Power Alleys (~2008-2016)... Messaging Leaders = Want to Change That KPCB INTERNET TRENDS 2016 | PAGE 108

Average Global Mobile User = ~33 Apps...12 Apps Used Daily... 80% of Time Spent in 3 Apps Day in Life of a Mobile User, 2016 Average # Apps Average Number Average Number of Time Spent on Most Commonly Installed on of Apps Used Apps Accounting for Phone (per Day) Used Apps Device* Daily 80%+ of App Usage Facebook USA 37 12 3 5 Hours Chrome YouTube Facebook Worldwide 33 12 3 4 Hours WhatsApp Chrome Source: SimilarWeb, 5/16. *Apps installed does not include pre-installed apps. Most commonly used apps includes preloads. KPCB INTERNET TRENDS 2016 | PAGE 109

Messaging Apps = Increasingly Becoming Second Home Screen... iOS Facebook Messenger Home Screen Inbox KPCB INTERNET TRENDS 2016 | PAGE 110

RE-IMAGINING HUMAN / COMPUTER INTERFACES – – VOICE – TRANSPORTATION

Re-Imagining Voice = A New Paradigm in Human-Computer Interaction KPCB INTERNET TRENDS 2016 | PAGE 112

Evolution of Basic Human-Computer Interaction Over ~2 Centuries = Innovations Every Decade Over Past 75 Years KPCB INTERNET TRENDS 2016 | PAGE 113

Human-Computer Interaction (1830s – 2015), USA = Touch 1.0  Touch 2.0  Touch 3.0  Voice Punch Cards for QWERTY Electromechanical Electronic Computer Paper Tape Reader Informatics Keyboard Computer (Z3) (ENIAC) (Harvard Mark I) 1832 1872 1941 1943 1944 Mainframe Computers Trackball Joystick Microcomputers Portable Computer (IBM SSEC) 1952 1967 (IBM Mark-8) (IBM 5100) 1948 1974 1975 Commercial Use of Commercial Use Commercial Use Touch + Camera - Voice on Mobile Voice on Connected / Window-Based GUI of Mouse of Mobile based Mobile (Siri) Ambient Devices (Xerox Star) (Apple Lisa) Computing Computing 2011 (Amazon Echo) 1981 1983 (PalmPilot) (iPhone 2G) 2014 1996 2007 Source: University of Calgary, “History of Computer Interfaces” (Saul Greenberg) KPCB INTERNET TRENDS 2016 | PAGE 114

Voice as Computing Interface = Why Now? KPCB INTERNET TRENDS 2016 | PAGE 115

Voice = Should Be Most Efficient Form of Computing Input Voice Interfaces – Voice Interfaces – Consumer Benefits Unique Qualities 1) Fast 1) Random Access vs. Humans can speak 150 vs. type 40 Hierarchical GUI words per minute, on average... Think Google Search vs. Yahoo! Directory... 2) Easy Convenient, hands-free, instant... 2) Low Cost + Small Footprint Requires microphone / speaker / 3) Personalized + Context- processor / connectivity – great for Driven / Keyboard Free Internet of Things... Ability to understand wide context of questions based on prior questions / 3) Requires Natural Language interactions / location / other semantics Recognition & Processing Source: Learn2Type.com, National Center for Voice and Speech, Steve Cheng, Global Product Lead for Voice Search, Google KPCB INTERNET TRENDS 2016 | PAGE 116

Person to Machine (P2M) Voice Interaction Adoption Keys = 99% Accuracy in Understanding & Meaning + Low Latency As speech recognition accuracy goes from say 95% to 99%, all of us in the room will go from barely using it today to using it all the time. Most people underestimate the difference between 95% and 99% accuracy – 99% is a game changer... No one wants to wait 10 seconds for a response. Accuracy, followed by latency, are the two key metrics for a production speech system... ANDREW NG, CHIEF SCIENTIST AT BAIDU Source: Andrew Ng, Chief Scientist, Baidu Note: P2M = person to machine. KPCB INTERNET TRENDS 2016 | PAGE 117

Machine Speech Recognition @ Human Level Recognition for... Voice Search in Low Noise Environment, per Google Next Frontier = Recognition in heavy background noise in far-field & across diverse speaker characteristics (accents, pitch...) Words Recognized by Machine (per Google), 1970 – 2016 10,000,000 1,000,000 ne @ ~90% hi accuracy c 100,000 a M @ ~70% d by 10,000 accuracy e z ogni 1,000 c e R ds 100 or W 10 1 1970 1980 1990 2000 2010 2016 Source: Johan Schalkwyk, Voice Technology and Research Lead, Google Note: For the English language. KPCB INTERNET TRENDS 2016 | PAGE 118

Voice Word Accuracy Rates Improving Rapidly... +90% Accuracy for Major Platforms Word Accuracy Rates by Platform*, 2012 – 2016 *Word accuracy rate definitions are unique to each company...see footnotes for more details 100% 90% 80% ) % ( 70% e at 60% R acy 50% r 40% ccu A 30% d r 20% o W 10% 0% Baidu1 Google2 Hound Voice Search3 (2012 - 2016) (2013 - 2015) & Assistant App (2012 - 2016) Source: Baidu, Google, VentureBeat, SoundHound Note: *Word Error Rate (WER) definitions are specific to each company. Word accuracy rate = 1 - WER. (1) Data shown is word accuracy rate on Mandarin speech recognition on one of Baidu's speech tasks. Real world mobile phone speech data is very noisy and hard for humans to transcribe. A 3.5% WER is better than what most native speakers can accomplish on this task. WER across different KPCB INTERNET TRENDS 2016 | PAGE datasets and languages are generally not comparable. (2) Data as of 5/15 and refers to recognition accuracy for English language. Word error rate is evaluated using real world search data which is extremely diverse and more error prone than typical human dialogue. (3) Data as of 1/16 and refers to recognition accuracy for English language. Word accuracy rate based on data collected from 119 thousands of speakers and real world queries with noise and accents.

Computing Interface... Evolving from Keyboards to Microphones & Keyboards = Still Early Innings KPCB INTERNET TRENDS 2016 | PAGE 120

Mobile Voice Assistant Usage = Rising Quickly... Primarily Driven By Technology Improvements % of Smartphone Owners Using Voice Assistants Voice Assistant Usage – Primary Reason for Annually, USA, 2013 – 2015 Change, % of Respondents, USA, 2014 – 2015 80% Software / technology has 32% improved 35% 65% 60% More aware of products via 32% 56% advertising / friends / family / other ways 30% s nt Need to use more because of 23% ponde 40% lifestyle / schedule 20% s e 30% 9% R More relevant services to meet l needs a 9% ot T 20% of 3% Don't know why % 4% 2015 Other (Please Specify) 1% 2014 0% 2% 2013 2014 2015 Source: Thrive Analytics, “Local Search Reports” 2013-2015 Note: Results highlighted in these charts are from the 2013, 2014, and/or 2015 Local Search surveys. These surveys were conducted via an online panel with representative sample sizes for the national KPCB INTERNET TRENDS 2016 | PAGE population in the US. There were 1,102, 2,058, and 2,125 US smartphone owners that completed the surveys in 2013, 2014 and 2015 respectively. 121

Google Voice Search Queries = Up >35x Since 2008 & >7x Since 2010, per Google Trends Google Trends imply queries associated with voice-related commands have risen >35x since 2008 after launch of iPhone & Google Voice Search Google Trends, Worldwide, 2008 – 2016 Navigate Home Call Mom Call Dad 2008 2009 2010 2011 2012 2013 2014 2015 2016 Source: Google Trends Note: Assume command-based queries are voice searches given lack of relevance for keyword-based search. Aggregate growth values determined using growth in Google Trends for three queries listed KPCB INTERNET TRENDS 2016 | PAGE above. 122

Baidu Voice = Input Growth >4x...Output >26x, Since Q2:14 Usage across all Baidu products growing rapidly...typing Chinese on small cellphone keyboard even more difficult than typing English...Text-to-Speech supplements speech recognition & key component of man-machine communications using voice Baidu Speech Recognition Daily Usage by API Calls, Baidu Text to Speech (TTS) Daily Usage by API Calls, 1 2 Global, 2014 – 2016 Global, 2014 – 2016 lls lls a a I C I C P P A A Q2:14 Q3:14 Q4:14 Q1:15 Q2:15 Q3:15 Q4:15 Q1:16 Q2:14 Q3:14 Q4:14 Q1:15 Q2:15 Q3:15 Q4:15 Q1:16 Source: Baidu Note: (1) Data shown is growth of speech recognition at Baidu, as measured by the number of API calls to Baidu's speech recognition system across time, from multiple products. Most of these API calls were for Mandarin speech recognition. (2) Data shown is growth of TTS (text to speech) at Baidu, in terms of the total number of API calls to Baidu's TTS system across time, from multiple products. Most of KPCB INTERNET TRENDS 2016 | PAGE these API calls were for Mandarin TTS. 123

Hound Voice Search & Assistant App = 6-8 Queries Across 4 Categories per User per Day Seeing 6-8 queries per active user per day among 100+ domains across 4 categories... Users most care about speed / accuracy / ability to follow up / ability to understand complex queries... Voice Query Breakdown – Observed Data on Hound App, USA, 2016 Local Fun & Information Entertainment 22% 21% Personal General Assistant Information 27% 30% Source: SoundHound Note: Based on most recent 30-days of user activity. Local information refers to queries about weather, restaurants, hotels, maps and navigation. Fun & entertainment refers to queries about music, movies, games, etc. General information refers to queries about facts, dictionary, sports, stocks, mortgages, nutrition, etc. Personal assistant refers to queries and commands about phone / communications, Uber KPCB INTERNET TRENDS 2016 | PAGE and transportation, flight status, calendars, timers, alarms, etc. 124

Voice = Gaining Search Share... USA Android @ 20%...Baidu @ 10%...Bing Taskbar @ 25% September 2014 May 2016 2020 Baidu – 1 in 10 queries Bing – 25% of In five years time at least 50% come through speech. searches performed on of all searches are going to be Windows 10 taskbar either through images or are voice searches per speech. Microsoft reps. Andrew Ng Chief Scientist, Baidu (9/14) June 2015 2015 May 2016 Siri – handles more Amazon Echo – Android – 1 in 5 than 1 billion requests fastest-selling speaker searches on mobile per week in 2015, @ for ~25% of app in USA are voice through speech. USA speaker market, searches & share is per 1010data. growing. Source: Baidu World 2014, Gigaom, Gadgets 360, 1010data, MediaPost, SearchEngineLand, Google I/O 2016, ComScore, Recode, Fast Company KPCB INTERNET TRENDS 2016 | PAGE 125

Voice as Computing Interface... Hands & Vision-Free = Expands Concept of ‘Always On’ KPCB INTERNET TRENDS 2016 | PAGE 126

Hands & Vision-Free Interaction = Top Reason to Use Voice...@ Home / In Car / On Go Primary Reasons for Using Voice, Primary Setting for Voice Usage, 1 2 USA, 2016 USA, 2016 Useful when hands / vision 61% Home occupied 43% Faster 30% results Difficulty typing Car 36% on certain 24% devices They're fun 22% / cool On the go 19% To avoid confusing 12% menus Other 1% Work 3% 0% 20% 40% 60% 80% 0% 10% 20% 30% 40% 50% Source: MindMeld “Intelligent Voice Assistants Research Report – Q1 2016” Note: Based on survey of n = 1,800 respondents who were smartphone users over the age of 18, half female half male, geographically distributed across the United States. (1) In response to the survey KPCB INTERNET TRENDS 2016 | PAGE question stating “Why do you use voice/search commands? Check all that apply.” (2) In response to the survey question stating “Where do you use voice features the most?” 127

Voice as Computing Interface... Platforms Being Built... Third Party Developers Moving Quickly KPCB INTERNET TRENDS 2016 | PAGE 128

Amazon Alexa Voice Platform Goal = Voice-Enable Devices = Mics for Home / Car / Mobiles... Alexa Voice Service – OEM / Developer Integrations (10+ integrations...) Home Car On Go (Various OEMs) (Ford Sync) (Lexi app) Ring Invoxia Philips Hue Ecobee Scout Security ToyMail Luma Alexa ‘Skills’ Kit Developers = ~950 Skills (5/16) vs. 14 Skills (9/15) Source: TechCrunch, Amazon Alexa, AFTVnews Image: Geekwire.com, Heylexi.com KPCB INTERNET TRENDS 2016 | PAGE Note: Amazon launched the Alexa Skills Kit for third-party developers in 6/15. 129

...Amazon Alexa Voice Platform Goal = Faster / Easier Shopping on Amazon Leveraging proliferation of microphones throughout house to reduce friction for making purchases... * 3x faster to shop using microphone than to navigate menus in mobile apps ... Amazon Echo Amazon Prime (~44MM USA Subscribers) Evolution of Shopping Amazon Amazon with Echo Echo Dot Echo Tap 1. Shopping Lists (2014) 2. Reorder past purchases by voice (2015) 3. Order new items – assuming you are fine with Amazon selecting exact item (2015) Source: Cowen & Company Internet Retail Tracker (3/16), Recode, MindMeld Image: Amazon.com, Gadgets-and-tech.com, Tomaltman.com, Techtimes.com, Venturebeat.com KPCB INTERNET TRENDS 2016 | PAGE Note: *Per MindMeld study comparing voice-enabled commerce to mobile commerce for the following task, “show me men’s black Adidas shoes for under $75” – takes ~7 seconds using voice compared to 130 ~3x longer navigating menus in an app.

~5% of Amazon USA Customers Own an Echo vs. 2% Y/Y... ~4MM Units Sold Since Launch (11/14), per CIRP ~4MM Amazon Echo devices have been sold in USA as of 3/16, with ~1MM sold in Q1:16, per CIRP estimates Amazon Customer Awareness of Amazon Amazon Customer Ownership of Amazon Echo, USA, Q1:15 – Q1:16 Devices, USA, Q1:16 70% 60% 61% 51% 60% 50% 50% 47% 40% ase 40% ase 34% B40% B er er30% 26% m 30% m o 30% o 22% st st u 20% u 20% C C f 20% f o o % % 10% 10% 6% 5% 0% 0% Q1:15 Q2:15 Q3:15 Q4:15 Q1:16 Prime Kindle Kindle Fire Echo None Fire Reader TV Source: Consumer Intelligence Research Partners (CIRP) Note: Amazon Echo limited launch occurred in 11/14 and wide-release occurred in 6/15. KPCB INTERNET TRENDS 2016 | PAGE 131

Computing Industry Inflection Points = Typically Only Obvious With Hindsight KPCB INTERNET TRENDS 2016 | PAGE 132

iPhone Sales May Have Peaked in 2015... While Amazon Echo Device Sales Beginning to Take Off? iOS Smartphone Unit Shipments, Estimated Amazon Echo Unit Shipments, Global, 2007 – 2016E USA, Q2:15 – Q1:16 250 200 ~1MM (MM) 150 (MM) ts ts n n e e m 100 m p p i i h h S S t t i i n n U 50 U 0 07 08 09 10 11 12 13 14 15 16E Q2:15 Q3:15 Q4:15 Q1:16 20 20 20 20 20 20 20 20 20 20 Source: Morgan Stanley Research (5/16), Consumer Intelligence Research Partners (CIRP), KPCB estimates Note: Apple unit shipments shown on a calendar-year basis. Amazon Echo limited launch occurred in 11/14 and wide-release launch occurred in 6/15. KPCB INTERNET TRENDS 2016 | PAGE 133

Re-Imagining Transportation = Another New Paradigm in Human-Computer Interaction... Cars KPCB INTERNET TRENDS 2016 | PAGE 134

Is it a Car...Is it a Computer?... Is it a Phone...Is it a Camera? Is it a Car...Is it a Computer? Source: Apple, Tesla KPCB INTERNET TRENDS 2016 | PAGE 135

...One Can... Lock / Monitor / Summon One’s Tesla from One’s Wrist Source: Tesla, The Verge, Redmond Pie KPCB INTERNET TRENDS 2016 | PAGE 136

Car Industry Evolution = Computerization Accelerating KPCB INTERNET TRENDS 2016 | PAGE 137

Car Computing Evolution Since Pre-1980s = Mechanical / Electrical  Simple Processors  Computers Pre-1980s Today = Complex Analog / Mechanical Computing Used switches / wiring to route Up to 100 Electronic Control feature controls to driver Units / car... Multiple bus networks per car (CAN / LIN / FlexRay / MOST)... 1980s (to Present) Drive by Wire... CAN Bus (Integrated Network) New regulatory standards drove need to monitor emissions in Today = Smart / real time, hence central Connected Cars computer Embedded / tethered connectivity... Big Tech = New Tier 1 auto 1990s (to Present) supplier OBD (On-Board (CarPlay / Android Auto)... Diagnostics) II Monitor / report engine performance; Required in all “The Box” USA cars post-1996 Tomorrow = Computers (Brooks & Bone) Go Mobile?... 1990s-2010s Central hub / decentralized Feature-Built Computing systems? + Early Connectivity LIDAR... Automatic cruise control... Vehicle-to-Vehicle (V2V) / Infotainment...Telematics... GPS Vehicle-to-Infrastructure (V2I) / / Mapping... 5G... Security software... rd Source: KPCB Green Investing Team, Darren Liccardo (DJI); Reilly Brennan (Stanford); Tom Denton, “Automobile Electrical and Electronics Systems, 3 Edition,” Oxford, UK: Tom Denton, 2004; Samuel DaCosta, Popular Mechanics, Techmor, US EPA, Elec-Intro.com, Autoweb, General Motors, Garmin, Evaluation Engineering, Digi-Key Electronics, Renesas, Jason Aldag and Jhaan Elker / Washington KPCB INTERNET TRENDS 2016 | PAGE Post, James Brooks / Richard Bone, Shareable 138

Car Automation Accuracy / Safety Improvements = Accelerating... Early Innings of Level 2 / Level 3 NHTSA – Automated Driving System Classifications L0 L1 L2 L3 L4 No Function- Combined Limited Full Automation Specific Function Self-Driving Self-Driving Automation Automation Automation Automation • Driver in complete and • Automation of one or • Automation of at least • Driver able to cede full • Vehicle can perform all sole control of primary more primary vehicle two primary vehicle control of all safety- safety-critical driving on vehicle controls (brake, control functions, but no control systems critical functions under and monitoring i steering, throttle, motive combination of systems working in unison certain conditions. functions during an pt power) at all times. working in unison Driver is expected to be entire trip i Systems with warning r available for occasional c technology (e.g. control, but with s forward collision sufficiently comfortable De warning) do not imply transition time automation • N/A • ABS • Tesla Autopilot • Google Car (manned • Google Car e • Cruise Control • GM Super Cruise prototype) l p • Electronic Stability (2017) m Control xa • Park Assist E • Since cars invented • 1990s – Today • 2010s • 2010s • ? e e (1760s) m m Ti a Fr Source: National Highway Traffic Safety Administration, “Policy on Automated Vehicle Deployment” (5/2013), Tesla, General Motors, Google, media reports KPCB INTERNET TRENDS 2016 | PAGE 139

Early Autonomous / ADAS Features Continue to Improve = Miles Driven Continue to Rise Google (Level 3 / 4 Autonomy) Tesla (Level 2 Autonomy) Source: Google, Tesla, Steve Jurvetson, EmTech Conference, The Verge KPCB INTERNET TRENDS 2016 | PAGE 140

Primary Approaches to Autonomous Vehicle Rollouts = All New or Assimilation...Traditional OEMs Taking Combined Approach All New = Assimilation = Top-Down, Fully Gradual Rollout / Autonomous Vehicles Mixed-Fleet Environments • Design & build vehicles from day one with • Roll out / upgrade autonomous features goal of full autonomy in current automotive context • Craft architectures / systems for end • Solves issue of integrating autonomy into product needs and with full fleet in mind existing asset base • Adapt testing environments to needs • Real-time, in-field updates & (individual city testing) improvements (Tesla over-the-air • Solves potentially dangerous middle layer software updates)...real-world learnings of semi-autonomy • Semi-autonomous stages require • Need very specific environments and potentially dangerous resumption of regulation to guide integration with driver control current system • OEM production cycles sometimes long, • Potentially difficult to scale which could cause innovation to remain slow • Key Example: • Key Example: Source: Google, Tesla, Morgan Stanley Research, Reilly Brennan (Stanford) KPCB INTERNET TRENDS 2016 | PAGE 141

Car Industry Evolution = Driven by Innovation... USA Led...USA Fell KPCB INTERNET TRENDS 2016 | PAGE 142

Car Industry Evolution, 1760s – Today = Driven by Innovation + Globalization Early Innovation Streamlining Modernization Re-Imagining Cars (1760s-1900s) = (1910s-1970s) = (1970s-2010s) = (Today) = European Inventions American Leadership Going Global / Mass Market USA Rising Again? 1768 = First Self-Propelled Road 1910s = Model T / 1960s = Ralph Nader / DARPA Challenge (2004, 2005, Vehicle (Cugnot, France) Assembly Line (Ford) Auto Safety 2007, 2012, 2013) = Autonomy Inflection Point? 1970s = Oil Crisis / 1876 = First 4-stroke cycle engine Emissions Focus (Otto, Germany) 1920s-1930s = Today = Car as Status Symbol... Roaring ‘20s / First Motels \ 1980s = Japanese Auto Takeover Begins... 1886 = First gas-powered, ‘production’ vehicle + (Benz, Germany) 1950s = Golden Age... Interstate Highway Act (1956)... 1990s – 2000s = 8 of Top 10 in Fortune 500 Industry Consolidation; in Cars or Oil (1960) Asia Rising; + 1888 = First four-wheeled USA Hybrid Fail (Prius Rise) electric car (Flocken, Germany) ? Late 2000s = Recession / Bankruptcies / Auto Bailouts Source: KPCB Green Investing Team, Reilly Brennan (Stanford), Piero Scaruffi, Inventors.About.com, International Energy Agency, Joe DeSousa, Popular Science, Franz Haag, Harry Shipler / Utah State Historical Society, National Archives, texasescapes.com, Federal Highway Administration, Matthew Brown, Forbes, Grossman Publishers, NY Times, Energy Transition, UVA Miller Center for Public Affairs, KPCB INTERNET TRENDS 2016 | PAGE The Detroit Bureau, SAIC Motor Corporation, Hyundai Motor Company, Kia Motors, Toyota Motor Corporation, DARPA, Chris Urmson / Carnegie Mellon, 143

Global Car Production Share = Rise & Decline of USA... Cars Produced in USA = 13% vs. 76% (1950)... Annual Light Vehicle Production, By Region, 1950 – 2014 100 90 ) M 80 M ( on 70 i t 60 oduc r 50 P e l c 40 hi e V 30 ght 20 Li 10 0 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 USA China Japan Western Europe Other Source: Wards Automotive, Morgan Stanley Research Note: Production measure represents all light vehicles manufactured within the given region (regardless of OEM home country). Light vehicles include passenger cars, sport utility vehicles and light trucks KPCB INTERNET TRENDS 2016 | PAGE (e.g. pickups). Data from 1950-1985 only available every 5 years. Largest “Other” constituents are South Korea, India and Mexico. 144

Detroit Population Tells Tale of USA Car Production = Down 65% from 1950 Peak @ 1.8MM Detroit Population, 1900 – 2015 2.0 1.8 1.6 1.4 (MM) n o 1.2 ti a l u 1.0 p o P t i 0.8 o tr e 0.6 D 0.4 0.2 0.0 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010 2015 Source: Southeast Michigan Council of Governments Note: Represents mid-year population. KPCB INTERNET TRENDS 2016 | PAGE 145

Car Industry = Innovation Accelerating in USA KPCB INTERNET TRENDS 2016 | PAGE 146

USA = Potential to be Global Hub of Auto Industry Again?... USA Has Many Key Components of Ecosystem 1) Incumbents – GM / Ford...Leading (2 of Top 10 Global) Auto Manufacturers 2) Attackers – Tesla... #1 Electric Vehicle Manufacturer 3) Systems / Components – Processors / GPUs (Nvidia...)...Sensors / LIDAR / Radar (Velodyne / Quanergy / Google...)...Connectivity (AT&T / Telogis / INRIX...)...Mapping (Google / Waze / Uber...)...Operating Systems (Google / Apple)...Other (Drivetrain / Power Electronics / Aerodynamics / Lightweighting / Etc...) 4) Autonomous Vehicles – Google / Tesla / Uber...Leadership in Development of Autonomous Vehicle Solutions 5) Mobility & Fleet Innovation – Uber / Lyft / Zendrive...Leadership in Ride Sharing Solutions / Infrastructure / Fleet Knowledge (Distribution via Mobile Devices / Recommended Traffic Flows) 6) Education / University Innovation – Stanford / Carnegie Mellon / Michigan / MIT / UC Berkeley...Leadership in STEM & Computer Science Education / Computer Vision / Robotics / Deep Learning / Automotive Engineering Source: KPCB Green Investing Team, Reilly Brennan (Stanford) KPCB INTERNET TRENDS 2016 | PAGE 147

...USA = Potential to be Global Hub of Auto Industry Again? USA Could Benefit from Creating Space in the Automotive Regulatory Framework to Foster Innovation 1) Federally Provided Guidance to States to Embrace Autonomy – Multiple legislative frameworks from individual states could impede autonomous innovation... 2) Flexibility of Regulation – Numerous approaches to solving autonomy challenge are likely to evolve simultaneously... regulation should not impede any single innovation approach... 3) Individual Cities / States Championing Autonomy – More testing locations / forward-leaning cities like Mountain View, CA / Austin, TX / Kirkland, WA / Metro Phoenix, AZ... 4) Comprehensive Safety Frameworks – Gov’t should have power to allow autonomous systems that demonstrate quantifiable safety improvements over current driver-vehicle combination... 5) Leaning Forward on Sharing (Car & Ride) – Regulators should work with rather than against sharing companies to craft policy as consumer demand illustrates need / interest in sharing... 6) Auto Cybersecurity – Connected cars face increased risk of cyber attacks...manufacturers & suppliers should keep consumer security / privacy as a key priority... 7) Next-Generation Franchise Laws – Semi-autonomous & autonomous cars are likely to change process of buying / servicing given ‘over the air’ nature of software downloads...USA could consider the EU ‘Block Exemption’ as model & allow consumers to service vehicles at either manufacturer-affiliated or independent locations Source: KPCB Green Investing Team, Reilly Brennan (Stanford), Google Note: EU Block Exemption details per European Commission. Testing locations represent Google autonomous car testing cities. KPCB INTERNET TRENDS 2016 | PAGE 148

Regulators = Typically Slow to Adapt to New Technologies Back in the Day When Horseless Carriage (Car) Came Along... Locomotive Act of 1865 – Jitneys (1914) Red Flag Act Ride-Sharing, ~100 Years Ago... Law Enacted in UK... 150K Jitney Rides / Day (1915) in LA, yet Horseless Carriages (Cars) Had to be Regulated Out of Existence by 1919... Preceded By Someone with Red Flag For Safety Purposes 157K Uber Rides / Day (2016) in LA... Source: Encyclopaedia Brittanica, dailybritain.wordpress.com, Travis Kalanick (Uber) TED Talk (3/16), Michigan State University Library, William B. Friedricks, “Henry E. Huntington and the Creation of Southern California,” Columbus, OH: Ohio State University Press, 1992 KPCB INTERNET TRENDS 2016 | PAGE 149

Global Perspective on Auto Industry Future – By Region, per Morgan Stanley Auto & Shared Mobility Research N. America – Some home field advantage on tech innovation & early application of shared mobility, but culture of private ownership and litigious USA judicial system may slow progress. China – Government focus on technology / environment, as well as quality of ride-sharing companies (esp. Didi), have driven strong early sharing adoption. Competing investment in public transit and impact of car ownership on social standing may impede full-scale adoption. India – Offers all key ingredients (rapid urbanization, limited public infrastructure, large millennial population, internet inflection point) for shared mobility leadership. Current market structure is likely to change as shared mobility gains dominance, so future remains unclear. Europe – Lack of homegrown tech champions coupled with power of OEMs (particularly Germans) and quality of European public transit may make adoption more difficult. High fuel costs and strong emissions standards may drive movement forward. Japan – Social implications of an aging population and policy support (given importance of a strong automotive industry) represent key advantages, but OEM buy-in to new paradigm is crucial, and R&D investment in tech arena lags somewhat behind other geographies. Korea – Strong technological culture, early political support and sharing-focused younger demographic leaves Korea relatively well positioned for move to shared mobility, though adoption remains in its infancy. Source: ‘Global Investment Implication of Auto 2.0,’ Morgan Stanley Research, 4/19/16, led by Adam Jonas KPCB INTERNET TRENDS 2016 | PAGE 150

Re-Imagining Transportation – Mobility also Being Re-Imagined KPCB INTERNET TRENDS 2016 | PAGE 151

Re-Imagining Automotive Industry = From Cars Produced to Miles Driven? We do believe the traditional ownership model is being disrupted...We’re going to see more change in the next five to ten years than we’ve seen in the last 50. MARY BARRA, GM CEO, 10/25/15 You could say there would be less vehicles sold, but we’re changing our business model to look at this as vehicle miles traveled...I could argue that with autonomous vehicles, the actual mileage on those vehicles will accumulate a lot more than a personally owned vehicle. MARK FIELDS, FORD CEO, 4/12/16 Source: Mary Barra (General Motors), Mark Fields (Ford), Wall Street Journal KPCB INTERNET TRENDS 2016 | PAGE 152

Car Ownership Costs (Money + Time) = High Car Ownership Costs = High $8,558 / Year, USA = Depreciation @ 44% / Fuel @ 15% / Finance + Fees @ 14% / Insurance @ 14% / Maintenance + Repair @ 9% Commuting Time = Significant 4.3 Hours per Week per Worker, Average (13% of Work Week, USA) Urban Auto Commuting Delays = Rising 42 Hours / Year / Urban Worker, USA (+2x in 30 Years), Equivalent to ~1.2 Extra Work Weeks / Year Millennials = Driving Differently Drivers License Usage Declining (Age 16-44) = @ 77% vs. 92% (1982, USA) Millennial Willingness to Car Share = @ ~50% (Asia-Pacific) / @ ~20% (North America) 46% of Millennials Expect Vehicle Technology to do Everything a Smartphone Can... Source: Ownership costs per AAA (4/16); Vehicle fees include license, taxes and registration. Commuting times per U.S. Census Bureau (2013) and include all transport options apart from walking and biking. Average USA work week per OECD Employment Outlook (7/15). Urban auto commuting delays per Texas A&M Transportation Institute / INRIX 2015 Mobility Scorecard (8/15); delays defined as extra time spent during the year traveling at congested rather than free-flow speeds by private vehicle drivers / passengers for 471 US urban areas. Driver’s license rates per University of Michigan KPCB INTERNET TRENDS 2016 | PAGE Transportation Research Institute / Federal Highway Administration (1/16). Car sharing statistics per Goldman Sachs Research (5/15). Millennial expectations per AutoTrader 2016 Cartech Impact Study 153 (9/15, n=1,012).

Efficiency Gain Potential from Ride & Car Sharing = High Cars = Underutilized Assets USA = 2.2 Cars / Household, ~20% of Households Have 3+ Cars, Cars Used ~4% of Time Vehicle Miles Traveled (VMT) = High Per Capita USA VMT Per Capita = 9K / +11x China (~850) / +48x India (~200) Parking Infrastructure = Lots of It ~19MM Parking Spaces in Los Angeles County (2010), +12MM since 1950 14% of Incorporated Land in Los Angeles County Allocated to Parking ~4 Estimated Parking Spots / Person in USA Energy Consumption by Light Vehicles = Significant ~500B Gallons of Fuel, Global (2014)... Source: Car utilization / penetration, VMT and energy consumption per “”Global Investment Implications of Auto 2.0”, Morgan Stanley Research (4/16); Los Angeles parking data per Mikhail Chester, Andrew Fraser, Juan Matute, Carolyn Flower and Ram Pendyala (2015) Parking Infrastructure: A Constraint on or Opportunity for Urban Redevelopment? A Study of Los Angeles Parking Supply and KPCB INTERNET TRENDS 2016 | PAGE Growth, Journal of the American Planning Association, 81:4, 268-286; parking spots / person per Stefan Heck / Stanford Precourt Institute of Energy. 154

Uber Platform / Network = Why Millions of Riders Have Taken >1B Rides Since 2009 Top Reasons Riders Choose Uber • 93% = Get to Destination Quickly • 87% = Safety • 84% = Too Much Alcohol to Drive • 83% = Save Money • 77% = Avoid Dealing with a Car • 65% = Option During Public Transit ‘Off' Hours Source: Berenson Strategy Group, Uber Note: Survey conducted in 11/15 across 801 riders who had taken at least one trip in the past 3 months in 24 USA Uber markets. KPCB INTERNET TRENDS 2016 | PAGE 155

Shared Private Rides Becoming Urban Mainstream = uberPOOL @ 20% of Global Uber Rides in <2 Years • 36 = Global UberPool Cities, +7x Y/Y • 100MM = UberPool Trips Since Launch (8/14) • 40% = UberPool as % of Total SF Rides • 30MM = China Rides / Month (in <1 Year) • >100K = Riders / Week in 11 Global Cities • 90MM = Vehicle Miles Traveled saved vs. UberX* • 1.8MM = Gallons of Gas Saved vs. UberX* Source: Uber. UberPool announced in August 2014. * Represents first 3 months of 2016. KPCB INTERNET TRENDS 2016 | PAGE 156

Re-Imagining Most Important Seat in Car = Back Seat, Again? Rolls Royce 10hp (1904) = Mercedes-Benz F 015 Designed for Rider ‘Luxury in Motion’ Concept (2015) = Déjà Vu? Commute Time = Significant Engagement / Entertainment Opportunity? 25 s 21 u 21 o i 19 r 20 a V , 15 13 13 th s n m 11 11 o r 10 M 10 / tfo a 6 r l e P s 5 U / s r 0 u o Facebook Spotify Commute Instagram Snapchat Pinterest Twitter Tinder LinkedIn H Time Source: Time Spent data per Cowen & Co. Research + SurveyMonkey (n = 2,059, 6/15, minutes / day spent across all cohorts and extrapolated to hours / month), except for Spotify (per Company). Commute data per US Census Bureau as of 2013; includes all modes of transportation apart from walking / biking. Assumes 25.9 minute one-way commute, assumed to be 5 days per week in both commute directions and 4.35 average weeks / month. Images per RREC / SWNS.com, Mercedes-Benz, carbodydesign.com KPCB INTERNET TRENDS 2016 | PAGE 157

Transportation Industry = Strap In for Next Few Decades KPCB INTERNET TRENDS 2016 | PAGE 158

Automotive Industry Golden Age, Take Two? What if a Car: • Is part of a network that provides a commuting service that comes to you? • Is the most advanced computing device you use? • In effect, is an on-demand cash generator, boosted by car / ride sharing? • Gives you safe driving pay-backs from your insurer? • Is safer, due to automation / reduced human error? • Drives itself? Parks itself? • Makes you want to commute? • Makes you more productive? KPCB INTERNET TRENDS 2016 | PAGE 159

CHINA = INTERNET LEADER ON MANY METRICS Hillhouse Capital* Provided China Section of Internet Trends, 2016 *Disclaimer – The information provided in the following slides is for informational and illustrative purposes only. No representation or warranty, express or implied, is given and no responsibility or liability is accepted by any person with respect to the accuracy, reliability, correctness or completeness of this Information or its contents or any oral or written communication in connection with it. A business relationship, arrangement, or contract by or among any of the businesses described herein may not exist at all and should not be implied or assumed from the information provided. The information provided herein by Hillhouse Capital does not constitute an offer to sell or a solicitation of an offer to buy, and may not be relied upon in connection with the purchase or sale of, any security or interest offered, sponsored, or managed by Hillhouse Capital or its affiliates.

China Macro = Robust Service-Driven Job & Income Growth... Despite Investment Slowdown KPCB INTERNET TRENDS 2016 | PAGE Hillhouse Capital 161

China Services Industries = 50%+ (& Rising) of China’s GDP & ~87% of GDP Growth China’s GDP by Sector, 1995 – 2015 Source: National Bureau of Statistics of China, CEIC, Goldman Sachs Global Investment Research. KPCB INTERNET TRENDS 2016 | PAGE Hillhouse Capital 162

China Services* Industries Job Growth = Accelerating... Offsetting Job Losses from Construction / Manufacturing / Agriculture China Annual Employment Change by Sector, 1995 – 2015 30 Agriculture Construction, Mining & Manufacturing ) M 20 Services* M ( Net Overall Employment Gain nge ha C 10 nt e m oy pl 0 m E l nnua A -10 -20 Source: National Bureau of Statistics of China, Wind Information. *Note: Services include wholesale, retail, transportation, storage, communication, accommodation, catering, finance, education, real estate and other services. KPCB INTERNET TRENDS 2016 | PAGE Hillhouse Capital 163

China Urban Disposable Income Per Capita = Continues to Grow @ Solid Rates China Urban Disposable Income per Capita & Y/Y % Growth, 1995 – 2015 $5,000 25% Urban Disposable Income per $) ( Capita a t $4,000 20% i Y/Y Growth ap C er $3,000 15% h e p t m w co o n Gr I /Y e l $2,000 10% Y sab o sp i D $1,000 5% an b r U $0 0% 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Source: CEIC, assume constant FX 1USD=6.5RMB. KPCB INTERNET TRENDS 2016 | PAGE Hillhouse Capital 164

China Internet @ 668MM Users = +6% vs. +7% Y/Y KPCB INTERNET TRENDS 2016 | PAGE Hillhouse Capital 165

China Internet Users = 668MM, +6% vs. 7% Y/Y...@ 49% Penetration China Internet Users, 2008 – 2015 800 40% 700 35% ) M 600 30% M s ( h ser 500 25% t w U o et 400 20% Gr n er % t n /Y I 300 15% Y a n i h 200 10% C 100 5% 0 0% 2008 2009 2010 2011 2012 2013 2014 2015 China Internet Users Y/Y Growth (%) Source: CNNIC. Internet user data is as of mid-year. KPCB INTERNET TRENDS 2016 | PAGE Hillhouse Capital 166

China Mobile Internet Usage Leaders... Tencent + Alibaba + Baidu = 71% of Mobile Time Spent Share of Mobile Time Spent, April 2016 WeChat Daily Mobile Time Spent = ~200 Minutes per User, Average QQ QQ Browser Tencent Tencent Video Tencent News Tencent Games QQ Music JD.com All Others QQ Reading 29% UCWeb Browser WeChat Taobao 35% Alibaba Weibo YouKu Video Momo Shuqi Novel QQ AliPay 10% AutoNavi Mobile Baidu Baidu iQiyi / PPS Video Baidu Browser Baidu Tieba 91 Desktop Baidu Maps All Other Note: Grouping of apps include strategic investments made by Tencent, Alibaba and Baidu. Only apps in top 50 by time spent share are called out. Source: QuestMobile, Trustdata, and Hillhouse estimates. KPCB INTERNET TRENDS 2016 | PAGE Hillhouse Capital 167

China Internet Traction = Advertising / Commerce / Travel / Financial Services Trends Often Compare Favorably to USA KPCB INTERNET TRENDS 2016 | PAGE Hillhouse Capital 168

China Online Advertising > TV (2015)... Online > 42% Total Ad Spend vs. 39% in USA China Annual Advertising Spend by Medium, 2007 – 2016E $50 50% ) B $ $40 40% nd nd ( pe S pe d S A $30 30% l ng a i ot s T i t r of e $20 20% dv % A t l ne r e nnua $10 10% nt A I na hi C $0 0% 2007 2008 2009 2010 2011 2012 2013 2014 2015E 2016E Internet TV Outdoor Print Radio Internet % of Total Source: GroupM China, April 2016 Forecast. Assume constant FX 1USD = 6.5RMB. USA advertising share data excludes out-of-home, video game, and cinema. KPCB INTERNET TRENDS 2016 | PAGE Hillhouse Capital 169

China E-Commerce Companies = Dominate Top Retailer Rankings vs. USA Peers... China Top 7 Retailers USA Top 7 Retailers by Revenue*, 2015 by Revenue*, 2015 Alibaba Wal-Mart JD.com CVS China Kroger Resources Suning Walgreens GOME Amazon Pure-Play Wal-Mart E-Commerce Target Auchan Costco Group $B $50B $100B $150B $B $100B $200B $300B $400B Source: Euromonitor. Note: *Revenue defined as retail value of goods excluding tax, and excluding certain transaction categories such as consumer-to-consumer, motor vehicles & auto parts, tickets, travel bookings, delivery foodservice, returns, and others, hence may differ from company disclosed total revenue or gross merchandise value figures. KPCB INTERNET TRENDS 2016 | PAGE Hillhouse Capital 170

...China E-Commerce Companies = Gaining Retail Share Faster than USA Peers... Share of China Total Retail Share of USA Total Retail Revenue*, 2010 – 2015 Revenue*, 2010 – 2015 7% 7% 6% 6% s s 5% e 5% e l l a a S Alibaba l Amazon.com S i l a i 4% t 4% a e t JD.com R eBay Re 3% A na S 3% U f Chi o 2% 2% of % % 1% 1% 0% 0% 2010 2011 2012 2013 2014 2015 2010 2011 2012 2013 2014 2015 Source: Euromonitor. Note: *Revenue defined as retail value of goods excluding tax, and excluding certain transaction categories such as consumer-to-consumer, motor vehicles & auto parts, tickets, travel bookings, delivery foodservice, returns, and others, hence may differ from company disclosed total revenue or gross merchandise value figures. KPCB INTERNET TRENDS 2016 | PAGE Hillhouse Capital 171

...China E-Commerce = Becoming More Social... 31% of WeChat Users Purchase via WeChat, +2x Y/Y % of WeChat Users Making Channels Through Which Users Made E-Commerce Purchase Through E-Commerce Purchase WeChat 40% s 31% ser Links to U 30% Other Apps JD Mall at 22% featured h within eC WeChat W 32% ed 20% ey v r 15% u S Group of Chats or % 10% Friends Circle WeChat 23% Public Accounts 0% 23% 2015 2016 Source: McKinsey’s 2016 China Digital Consumer Survey Report. KPCB INTERNET TRENDS 2016 | PAGE Hillhouse Capital 172

China Travel...Ctrip = Expansive One-Stop-Shop for Travelers... Priceline App (USA) Ctrip App (China) Hotel B&B, Hostel Transport Train / Bus / Ferry Ticket Destination Tour Guide Attraction Portable Wi-Fi for Roaming Restaurant Travel Visa / Shopping / Insurance Currency Conversion 24/7 Customer Service Source: Priceline, Ctrip. KPCB INTERNET TRENDS 2016 | PAGE Hillhouse Capital 173

...China Outbound Travel Penetration @ Inflection Point = Already World’s Biggest Outbound Tourism Spender Outbound Departures as % of Top 10 Outbound Tourism Population, 1970 – 2015 Spending Country, 2014 35% China $165B 30% China USA $146B on i t Japan a Germany $107B opul 25% S. Korea P UK $80B of 20% % France $59B s a 15% e Russia $55B ur t r 10% pa Canada $34B e D 5% Australia $32B bound 0% Brazil $30B ut O 70 73 76 79 82 85 88 91 94 97 00 03 06 09 12 15 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 Italy $29B Source: CLSA, World Bank. KPCB INTERNET TRENDS 2016 | PAGE Hillhouse Capital 174

China Smartphone-Based Payment Solutions = High Engagement Estimated Monthly Payment Transactions per User WeChat Payment USA Debit Card AliPay USA Credit Card 0 10 20 30 40 50 60 Source: US debit and credit card data defined as number of payments (including online and offline) a month per active general-purpose card. Active cards are those used to make at least KPCB INTERNET TRENDS 2016 | PAGE one purchase or bill payment in a month. Data per 2013 Federal Reserve Payments Study. AliPay / WeChat Pay stats per Hillhouse estimates. WeChat data includes peer-to-peer Hillhouse Capital payments such as virtual Red Envelopes. 175

WeChat Chinese New Year Payments = 8B Virtual Red Envelopes Sent, + 8x Y/Y... WeChat Virtual Red Envelopes Sent – Chinese New Years Eve, 2014 – 2016 8 8B ) B ( nt e 6 S s ope l e nv 4 d E e R l ua t r i V 2 of # 1B 20MM 0 2014 2015 2016 Source: Tencent. KPCB INTERNET TRENDS 2016 | PAGE Hillhouse Capital 176

...WeChat Payments = Can Drive Merchant Loyalty & CRM Source: 86 Research. KPCB INTERNET TRENDS 2016 | PAGE Hillhouse Capital 177

Ant Financial (~$60B Valuation*) = Leveraging Alibaba AliPay Scale... Building China Financial Services One-Stop-Shop Savings / MoneyMarket SMB Lending Funds $100B+ 260MM+ Users Cumulative Loans $150B+ AUM Payment 450MM+ AliPay Users $1+ Trillion Payment Volume in 2015 Consumer Loan / Instant Credit Credit Bureau / 50MM+ Online Insurance Cumulative / P2P Lending... Consumer Loan Users Source: Media reports, Ant Financial. *Financing in 4/16 KPCB INTERNET TRENDS 2016 | PAGE Hillhouse Capital 178

China Internet Emerging Momentum = On-Demand KPCB INTERNET TRENDS 2016 | PAGE Hillhouse Capital 179

China On-Demand Transportation = Global Leader... 4B+ Annualized Trips (+4x Y/Y...~70% Global Share) Annualized Global On-Demand Transportation Trip Volume by Region, Q1:13 – Q1:16 ~25MM ~750MM ~1.7B ~6.3B Annualized 30x Y/Y 2.3x Y/Y 3.7x Y/Y Trip Volume China N. America EMEA India SE Asia ROW Q1:13 Q1:14 Q1:15 Q1:16 Source: Hillhouse Capital estimates, include on-demand taxi, private for-hire vehicles, as well as on-demand for-hire motorbike trips booked through smartphone apps. KPCB INTERNET TRENDS 2016 | PAGE Hillhouse Capital 180

China On-Demand Transportation... China Cities = Fastest Global Growers Monthly Trips Since Inception, Uber China vs. Rest of World Source: Uber China chart per leaked CEO letter to investors in China in June 2015, third-party press releases. KPCB INTERNET TRENDS 2016 | PAGE Hillhouse Capital 181

PUBLIC / PRIVATE COMPANY DATA

Impact of Internet = Extraordinary & Broad But, in Many Ways... It’s Just Beginning KPCB INTERNET TRENDS 2016 | PAGE 183

Internet-Related Dislocations = Long-Time in Making...Still Early Stage Cord-Cutting Impacts Earnings for Traditional Media Companies... E-Commerce Impacts Revenue Growth for Traditional Retailers Media Retail Market Cap 2006 2016* Market Cap 1997 2016* Viacom $33B $18B Wal-Mart $69B $222B Netflix $1.4B $44B Amazon.com $400MM $341B Revenue 2006 2015* Revenue 1997 2015* Viacom $11B $13B Wal-Mart $118B $482B (+19% Y/Y) (-6% Y/Y) (+12% Y/Y) (-1% Y/Y) Netflix $1B $7B Amazon.com $148MM $107B (+46% Y/Y) (+23% Y/Y) (+9.4x Y/Y) (+20% Y/Y) Source: CapIQ, Public Filings * 2015 revenue for all companies reflects CY2015. Current market caps as of 5/31/16. Historical market caps for Wal-Mart / Amazon shown as of date of Amazon IPO (5/15/1997). Historical market caps for KPCB INTERNET TRENDS 2016 | PAGE Viacom / Netflix shown as of date of CBS spinoff from Viacom (1/3/2006). 184

Current Generation of Internet Leaders = Growing Faster than Previous Generation Marketplaces Commerce Gross Merchandise Value (GMV), Time Shifted Gross Merchandise Value (GMV), Time Shifted $20 Alibaba vs. eBay vs. Airbnb vs. Uber $200 Amazon.com vs. JD.com $15 Alibaba / Taobao JD.com ) eBay ) $150 B B Amazon.com $ $ V ($10 Airbnb V ( $100 M Uber M G $5 G $50 $0 $0 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 9 1011121314151617181920 Years Since Launch (T+) Years Since Launch (T+) Enterprise Est. Quarterly Revenue ($MM), Time Shifted Salesforce vs. Slack ) Salesforce M Slack $M e ( u en ev R 1 2 3 4 5 6 7 8 9 10 11 12 13 Marketplaces Source: Company data, Morgan Stanley Research. eBay founded in 1995. Amazon founded in 1995. Alibaba.com founded in 1999 as B2B portal connecting Chinese manufacturers and overseas buyers. Uber launched 2009, gave first ride in 2010. Airbnb founded in 2008.. KPCB INTERNET TRENDS 2016 | PAGE Commerce Source: Publicly available company data, Morgan Stanley Research. JD.com launched B2C shipments in 2004, founded 1998 as an online magneto-optical store. Amazon founded in 1995. 185 Enterprise Source: Slack. Graph starting point based on similar est. revenue figures. Salesforce quarterly revenue approximated from publicly disclosed annual GAAP revenues.

Internet Leaders = Getting Bigger...Staying Aggressive KPCB INTERNET TRENDS 2016 | PAGE 186

Global Internet Market Leaders = Apple / Google / Amazon / Facebook / Tencent / Alibaba...Flush with Cash...Private Companies Well Represented Rank Company Region Current Market Q1:16 2015 Value ($B) Cash ($B) Revenue ($B) 1 Apple USA $547 $233 $235 2 Google / Alphabet USA 510 79 75 3 Amazon USA 341 16 107 4 Facebook USA 340 21 18 5 Tencent China 206 14 16 6 Alibaba China 205 18 15 7 Priceline USA 63 11 9 8 Uber USA 63 -- -- 9 Baidu China 62 11 10 10 Ant Financial China 60 -- -- 11 Salesforce.com USA 57 4 7 12 Xiaomi China 46 -- -- 13 Paypal USA 46 6 9 14 Netflix USA 44 2 7 15 Yahoo! USA 36 10 5 16 JD.com China 34 5 28 17 eBay USA 28 11 9 18 Airbnb USA 26 -- -- 19 Yahoo! Japan Japan 26 5 5 20 Didi Kuaidi China 25 -- -- Total $2,752 $447* $554* Source: CapIQ, CB Insights, Wall Street Journal, media reports. Market value data as of 5/31/16. * Includes only public companies. Note: For public companies, colors denote current market value relative to Y/Y market value. Green = higher. Red = lower. Purple = newly public within last 12 months (applied here to both eBay and Paypal given Paypal spinoff on 7/20/15). Yellow = private companies, where market value represents latest publicly announced valuation. Ant Financial and Didi Kuaidi valuation per latest media reports as of KPCB INTERNET TRENDS 2016 | PAGE 5/2016. Ant Financial treated separately from Alibaba as Alibaba retains no control of Ant and will receive a capped lump sum payment in the event of an Ant liquidity event. Cash includes cash and 187 equivalents and short-term marketable securities plus long-term marketable securities where deemed liquid.

Traditional Industry Incumbents = Active in Acquisitions / Investments KPCB INTERNET TRENDS 2016 | PAGE 188

Incumbents = Increasingly Betting on Technology Companies to Fuel Growth... Non-Tech Acquisitions of Tech Companies +2.6x Since 2012 Select Acquisitions by Non-Tech Incumbents • Auto Consortia / Nokia Here • Liberty Interactive / Zulily • Avis / Zipcar • Monsanto / Climate • AxelSpringer / Business Corporation Tech Acquisitions by Non-Tech Corporate Buyers Insider • Neiman Marcus / • Disney / Maker Studios, Mytheresa.com Volume ($B) Playdom • Nordstrom / HauteLook • Disney + Fox + • Northwestern Mutual / NBCUniversal / Hulu Learnvest • First Data / Perka, Clover • Staples / Runa $28 • Ford / Livio • Target / DermStore.com • General Motors / Cruise • Under Armour / Automation MapMyFitness, MyFitnessPal $21 • Hudson Bay / Gilt Groupe • Walmart / Kosmix $19 Select Investments by Non-Tech Incumbents $11 • American Express / Concur • Lowes / Porch • Citi / Ayasdi, Betterment • NBCUniversal / BuzzFeed, • Coca-Cola / OneWeb Vox Media • Ford / Pivotal • Nikkei / Evernote • Fox Sports / DraftKings • Turner Sports / FanDuel 2012 2013 2014 2015 • General Motors / Lyft • USAA / TRUECar • Goldman Sachs / Dataminr, • Visa / Square Kensho, Symphony • Whole Foods / Instacart • J.P. Morgan / Prosper Marketplace Source: Morgan Stanley, CapitalIQ, Thomson Reuters Note: Includes technology targets >$100MM in value. KPCB INTERNET TRENDS 2016 | PAGE 189

Global Technology Financings = Solid Trends in Private Financings... Only 2 Tech IPOs 2016YTD Source: Morgan Stanley, Thomson Reuters Note: YTD Tech IPOs include SecureWorks and Acacia Communications. KPCB INTERNET TRENDS 2016 | PAGE 190

Global Technology Public + Private Financing Volume = Solid Relative to History Global US-Listed Technology IPO Issuance and Global Technology Venture Capital Financing, 1990 – 2016YTD $200 March 10, 2000 = July 20, 2015 = ) NASDAQ @ 5,049 Technology Market Peak, B NASDAQ @ 5,219 $ Technology IPO Volume ( e ($B) nd um Technology Private $157 a ol $150 O V Financing Volume ($B) P I ng i NASDAQ ogy nc $107 na $96 Fi $100 hnol $89 e c t e a T v i l r P $58 $50 $48 nnua ogy $50 $42 $48 $44 A $40 $28 $28 $36 $36 $34 $33 hnol $26 $22 $25 c $19 e $14 T $3 $3 $8 $7 $5 $0 VC Funding per $3 $3 $2 $5 $4 $4 $5 $5 $6 $8 $14 $18 $11 $8 $8 $9 $8 $9 $8 $9 $7 $7 $10 $8 $9 $13 $15 $16 Company ($MM) *Facebook ($16B IPO) = 75% of 2012 IPO $ value. **Alibaba ($25B IPO) = 69% of 2014 IPO $ value. Source: Thomson ONE, 2016YTD as of 5/26/16. VC Funding per Company ($MM) calculated as total venture financing per year divided by number of companies receiving venture financing. KPCB INTERNET TRENDS 2016 | PAGE Morgan Stanley Equity Capital Markets, 2016YTD as of 5/26/16. All global U.S.-listed technology IPOs over $30MM, data per Dealogic, Bloomberg, & Capital IQ. 191

There are pockets of Internet company overvaluation but there are also pockets of undervaluation... Very few companies will win – those that do – can win big... Over time, best rule of thumb for valuing companies = value is present value of future cash flows. KPCB INTERNET TRENDS 2016 | PAGE 192

DATA AS A PLATFORM / DATA PRIVACY CREATED BY KPCB PARTNERS TED SCHLEIN / ALEX KURLAND

Data as a Platform KPCB INTERNET TRENDS 2016 | PAGE 194

Global Data Growth Rising Fast = +50% CAGR since 2010... Data Infrastructure Costs Falling Fast = -20% CAGR Data in Digital Universe vs. Data Storage Costs, 2010 – 2015 10B $0.20 8B ge a $0.15 a at or D 6B t f S o of es B t G y ab 4B r et $0.10 pe P t os 2B C 0B $0.05 2010 2011 2012 2013 2014 2015 Data in Digital Universe (Petabytes) Storage Costs ($/GB) Source: IDC, May 2016. KPCB INTERNET TRENDS 2016 | PAGE 195

Data Generators = Increasing Rapidly Source: Apple, DJI, Waze, Tesla, Microsoft, Ring, Fitbit, B & H Foto & Electronics. KPCB INTERNET TRENDS 2016 | PAGE 196

Data = A New Growth Platform... Powering New Services / Systems / Apps The Large investments in fiber optic & last-mile cables created h t Network connectivity that facilitated the early Internet growth ow r G t ne r e Optimizing the network with software became far more capital nt I The efficient than additional capex buildouts...ultimately resulting in l Software the creation of pervasive networks (siloed data centers  oba l AWS)...& then pervasive software (Siebel  Salesforce) G or f ge a Emergence of pervasive software created the need to optimize r The e v Infrastructure the performance of the network & store extraordinary amounts Le of data at extremely low prices of s e c our Next Big Wave = Leveraging this unlimited connectivity & S The storage to collect / aggregate / correlate / interpret all of this data Data to improve people’s lives & enable enterprises to operate more efficiently Source: Adam Ghetti, Ionic Security; Ted Schlein, KPCB. KPCB INTERNET TRENDS 2016 | PAGE 197

Evolution of the Data Platform, 1990 – 2016 FIRST WAVE SECOND WAVE THIRD WAVE Constrained Data... Data Explosion / Chaos... Mass Data Intelligence... Monolithic Systems, Decentralized Systems, Pervasive Systems, Expensive Storage, Cheap Storage, Big/Fast Storage, Data for Targeted Use Cases Big Data Everywhere Data Instruments the Business BUSINESS VISUALIZATION DEPARTMENTAL INTELLIGENCE (BI) APPLICATIONS CLOUD BI Business Objects, Gainsight, Datadog, e Cognos, MicroStrategy InsideSales r a ftw CACHING o ORGANIZATION-WIDE S DATA INTEGRATION PREP / WRANGLING ANALYTICS PLATFORMS Informatica Evolution ETL Revolution Looker, Domo, Anaplan Breaking Apart Data Integrated Data Bottleneck INFRASTRCUTURE- into Everything DATA-CENTRIC y t CENTRIC SECURITY & i DATA INTEGRITY SECURITY & r MANAGEMENT MANAGEMENT Microsoft, Oracle ecu Palo Alto Networks, S FireEye Ionic Security, Tanium e r Age of Big Data tu Age of Big/Fast c u Age of Oracle, Sybase Hadoop, Teradata, tr Redshift, BigQuery, s Netezza, NetApp, EMC, a Greenplum Spark, Presto fr n I Source: Looker, Ionic Security, KPCB. KPCB INTERNET TRENDS 2016 | PAGE 198

Data is moving from something you use outside the workstream to becoming a part of the business app itself. It’s how the new knowledge worker is actually performing their job. FRANK BIEN, CEO OF LOOKER, 2016 KPCB INTERNET TRENDS 2016 | PAGE 199

Data as a Platform – A Few Companies Utilizing Analytics to Improve Business Efficiency... KPCB INTERNET TRENDS 2016 | PAGE 200

Data Analytics as a Platform = Looker THEN NOW Complex Tools Operated by Data Analysts, Looker Chaos of Data Silos Across the Company Data analytics platform built for both data analysts & non-technical business users that can scale throughout organizations Source: Looker. KPCB INTERNET TRENDS 2016 | PAGE 201

Customer Data & Relationship Intelligence as a Platform = SalesforceIQ THEN NOW Difficult to Customize, Lack of SalesforceIQ Automated Customer Insights CRM solution that helps businesses build stronger customer relationships by analyzing data & patterns to identify opportunities. Source: Bomgar Corporation, Salesforce. KPCB INTERNET TRENDS 2016 | PAGE 202

Data Mapping as a Platform = Mapbox THEN NOW Difficult & Expensive to Collect Data... Mapbox Limited In-App Digital Map Usage Worldwide maps crowdsourced by a community of smartphone users whose mobile navigation data facilitates real-time updates to the platform KPCB INTERNET TRENDS 2016 | PAGE Source: Forbes; Technical.ly; Philadelphia Police Department; Mapbox. 203

Cloud Data Monitoring as a Platform = Datadog THEN NOW Expensive & Clunky Point Solutions, Datadog Lengthy Implementation Cycles, Only Used by System Administrators Cloud monitoring platform for both System Administrators & Developers that automatically integrates 100+ sources in real-time to represent hundreds of thousands of cloud instances Source: IBM; Datadog. KPCB INTERNET TRENDS 2016 | PAGE 204

Data Security & Management as a Platform = Ionic Security THEN NOW Securing Infrastructure to Ionic Security Keep Data Safe Distributed data protection & management platform that has processed tens of billions of API requests to enable customers to secure & control their data Source: www.teach-ict.com; Ionic Security. KPCB INTERNET TRENDS 2016 | PAGE 205

As Data Explodes... Data Security Concerns Explode KPCB INTERNET TRENDS 2016 | PAGE 206

Data Privacy Debate – Major Events, 2013 – 2016 Edward Snowden (Jun-13) Burr-Feinstein Anti-Encryption Bill (Apr-16) Former CIA contractor leaked classified information to media Proposed law that would require technology companies & phone about internet & phone surveillance by USA intelligence. manufacturers to decrypt customer data at a court’s request. Apple vs. FBI (Feb-16) Microsoft Lawsuit (Apr 16) FBI claimed it needed Apple to provide access to an iPhone Files lawsuit for right to be able to tell customers when law owned by a man who committed a mass shooting in San enforcement officials request their emails & other data. Bernardino, CA, so that the agency could recover information for its investigation. Request was denied by a federal judge in New York. . WhatsApp’s Default End-to-End Encryption Apple Hires Data Security Expert (Apr-16) (May-16) WhatsApp implements end-to-end encryption as default Jon Callas, who co-founded several well-respected secure setting to protect communications of their 1B monthly active communications companies including PGP Corp, Silent Circle and users worldwide. Blackphone, rejoins Apple (he was also an employee in the 1990s and again between 2009 and 2011, when he designed an encryption system to protect data stored on a Macintosh computer). Source: NY Times, CNBC, Reuters, Time, Washington Post, WhatsApp. KPCB INTERNET TRENDS 2016 | PAGE 207

Cybercrime = Widespread Borderless Threat… ~4 Billion Data Records Breached Globally Since 2013 Records Breached, Billions of Individual Records, Global, 2013 – 2015 3B ) B ( 2B es Includes 1.2B unique records breached by a Russian ch CyberGang called CyberVor. a e r B f o # 1B 0B 2013 2014* 2015 Source: Breach Level Index; IBM; Govtech Note: *Includes 1.2B unique records breached by a Russian CyberGang called CyberVor. KPCB INTERNET TRENDS 2016 | PAGE 208

Consumer Data Privacy Concerns Rising Rapidly How Concerned are You About Data Privacy & How Companies Use Customer Data? 4% 45% Are more worried about their 46% 50% Online privacy than one year ago 74% Have limited their online activity Very Concerned in the last year due to privacy concerns Somewhat Concerned Not Concerned Source: Gigya “The 2015 State of Consumer Privacy & Personalization” report, US respondents, n = 2,000; TRUSTe / National Cyber Security Alliance Consumer Privacy Survey – US, 2016. KPCB INTERNET TRENDS 2016 | PAGE 209

Consumers’ Top Privacy Concerns = Data Selling / Storage / Access / Being Identified Individually... Rate Level of Privacy Concerns Across Each of the Following Ways Companies Interact with Personal Data, n = 2,062 (These percentages reflect all respondents who rated their privacy concerns on a 1-5 scale, with 5 = Extremely Concerned, 4 = Very Concerned, etc.) If / Where they sell my data 78% Where they keep my data 73% How they identify me as an individual 68% How long they have my data 67% Who sees and analyzes the data 67% How a company gets my data 66% When and how I opted into sharing 61% How they use data to personalize marketing 59% How they use data to provide customer support 54% How they use data to improve or innovate 53% How they identify me as a group 52% Source: Altimeter Group, “Consumer Perceptions in the Internet of Things”, 2015. n = 2,062 respondents. KPCB INTERNET TRENDS 2016 | PAGE 210

...Do People Care About Privacy... Or Do They Care About Who Has Their Data? Amazon Echo Google Gboard The Echo’s Alexa Voice Service listens to all Integrated keyboard for iOS devices that had speech in default mode an estimated 500K+ downloads within the first week of launch Source: Amazon, Google, App Annie. KPCB INTERNET TRENDS 2016 | PAGE 211

In the tangible world, physical limitations prevent the broad abuse of the law... Should the same laws automatically apply to the digital world where a few lines of code can unlock someone’s entire life? ADAM GHETTI, FOUNDER & CEO OF IONIC SECURITY, 2016 KPCB INTERNET TRENDS 2016 | PAGE 212

Disclosure This presentation has been compiled for informational purposes only and should not be construed as a solicitation or an offer to buy or sell securities in any entity. The presentation relies on data and insights from a wide range of sources, including public and private companies, market research firms and government agencies. We cite specific sources where data are public; the presentation is also informed by non-public information and insights. We publish the Internet Trends report on an annual basis, but on occasion will highlight new insights. We will post any updates, revisions, or clarifications on the KPCB website. KPCB is a venture capital firm that owns significant equity positions in certain of the companies referenced in this presentation, including those at www.kpcb.com/companies. KPCB INTERNET TRENDS 2016 | PAGE 213