Editor’s note: The following is an unedited transcript from the 2018 Code Conference. It may have spelling or grammatical errors that will be corrected at a later date.
Mary Meeker: Thank you, guys. Good to be here. Thank you all, I should say.
I’m going to run through this presentation and I want to say a couple things. First, I want to thank Ansel and Michael who, and a lot of the folks at Kleiner Perkins who really help put this together. This is certainly not a one person show and a lot of people do a lot of heavy lifting. In addition, we have a section on China. Compiled by the folks at Hillhouse Capital specifically Liang Wu. This is a slide I’m not going to read. As those of you who have seen this before, this is a presentation that is meant to be read, not presented. This provides a lot of context on what we’re doing and the presentation is available at KleinerPerkins.com. I encourage you to read it because I will go through this stuff super fast.
This is the compilation of things we’re going to address. There are a few things we won’t get through. We won’t get through the advertising section, the Enterprise software stuff and the comments on USAA immigration, which the Senator addressed in a very thoughtful way.
I’m going to start out with internet devices and users growth continues to slow. Global smartphone new smartphone unit shipments had 0% growth in 2017 versus 2% growth in 2016. Internet users slowing growth of 7% versus 12% growth in the previous year. Global internet users of 3.6 billion surpassed half the world’s population in 2018. The reality of all for the business people in the room, when you get to a market, when you get to 50% penetration, new growth becomes a lot harder to find. That’s where the industry is at a really high level. That said, internet usage remains pretty solid, up 4% year end year with some U.S. data. It’s not deduped so there’s a lot of multitasking going on. A lot of people ask the question about internet usage, how much is too much? Our view is it depends on how the time is spent. One of the things I feel really strongly about is there’s a lot of innovation and there’s a lot of competition. That’s driving a lot of product improvements, a lot of usefulness and a lot of usage. Also, a lot of scrutiny.
We put together a list of the key areas where we think there’s a lot of innovation going on and there’s certainly a lot of growth. I’m going to zip through these really quickly.
Devices are getting better, faster and cheaper. Access is rising dramatically. This looks at the expansion of wifi networks around the world. Products are getting a lot easier to use and they’re becoming much more pervasive. Digital payments, digital reach is expanding. The portion of people around the wold that make payments and any given day, in a digital fashion, is quite high. With payments friction is declining, products like messengers and also, mobile payments are rising dramatically and digital currencies are emerging. This looks at the coin based user growth. Local online connections are driven by … Offline connections are driven by online network effects. This looks at Nextdoor’s neighborhood growth in the U.S. In messaging, extensibility is expanding. What used to be Q Q is now, a multidimensional series of types of products in the messenger area. With video, mobile adaption is rising very quickly and new content types are emerging. With voice, we’ve hit technology lift off with word accuracy and we’ve certainly product lift off with Amazon Echo’s install base estimated to be around 30 million plus.
The last theory to focus on in innovation competition is personalization. With personalization, data improves engagement in experiences and drives growth and scrutiny. Personal collective data provides better experiences for consumers. They’re 2.2 billion Facebooks, 200 million Pinterests, 170 million Spotifies and 125 million Netflixes. People putting their data into these products to make their experiences better and then there’s the collective data of many other users that effect a lot of real time products, whether it’s Waze or SnapMap or NextDoor, or Uber Pool. This all creates a privacy paradox. We tried to simplify it into really three sentences. Internet companies are making low price services better in part from user data. Internet users are increasing their time on internet services based on perceived value. Regulators want to ensure data is not used improperly and not all regulators think about this in the same way.
When we look at Facebook, the company is experiencing a higher engagement on the product and that drives monetization for the company and it drives investment in product improvements. When you have rising monetization, rising growth and rising data collection, it drives a lot of regulatory scrutiny whether it’s related to data privacy, competition or safety in content. For the internet companies, it’s key to understand in all this stuff and it’s very complex. We’re in the middle of it all right now and will be for a long time to come. It’s important to understand the unintended consequences of the products. This is Mark Zuckerberg’s quote in April. He said, “We did not focus enough on preventing abuse and thinking through how people could do these tools to do harm as well”. The unintended consequences applied to regulators as well. This is a comment from Bloomberg Opinion editorial as it comes into force Europe with GDPR should be mindful of unintended consequences and open to change when things go wrong.
Well, it’s crucial to manage for unintended consequences. It’s also irresponsible to stop innovation and progress, especially in a world where there are a lot of countries that are doing different things.
I want to drill down on the U.S. internet leaders. They’ve been aggressive and forward thinking investors for years. This looks at private money into public companies and IPO’s of companies over the last … beyond the last couple of decades, but money in has been quite, quite significant. Tech companies have risen to 25% of the market cap of the MSCI. That’s been a steady increase that tech companies account for six of the top fifteen RND and cap expenders in the U.S with Amazon, Google Alphabet, Intel, Apple and Microsoft in the top five slots and these companies, including Facebook, are the fastest growing spenders as well on that list. If we look at tech companies versus other industries in the United States, tech companies are the largest fastest growing RND and cap expenders. The blue line looks at attack a 9% combined average growth rate in RND and cap expending over the last ten years. Healthcare at 4% and discretionary at 0%. It’s just interesting to compare what the growth rates are in the different industries.
Tech companies have RND and cap x’s also rising as a percent of revenue at 18% versus 13% in 2007. A lot of competition, a lot of RND spending and cap expending driving a lot of innovation and growth.
I’m going to switch to eCommerce. The transformation continues to accelerate. eCommerce revenue was up 16% in 2017 in the U.S. versus 14% the previous year. Shared gains versus traditional retail continued to rise at 13% for eCommerce. Amazon.com, eCommerce share gains continue at 28% versus 20% in 2013. eCommerce evolving in scaling. When we think about eCommerce today, it’s often mobile interactive personalized in the feed, in the inbox and also often at the front door. We wanted to look at the stack of eCommerce just to give you a sense of some of the numbers in trends around how people are building stores and conducting commerce online. Often times offline merchants want to set up a payment system. A lot of times they’ll start a company like Square. Square has 2.8 million active sellers on its platform. We estimate a lot of times when people want to start and develop an online store, they’ll go to a service like Shopify. About six hundred thousand active merchants on their platform. Integrated payment systems strike. Integrate fraud protection, integrate purchase financing, integrate customer support, find customers and get products to the customers.
This industry, for the most part, has emerged over the course of the last five years and there’s no better indication in our view of how vibrant it is and how intensely competitive it is that Shopify is set up a store front exchange where you can buy and sell online stores that have been created on the platform.
I want to spend a little bit of time talking about how people find products and how that evolves. Search leads on the internet, most people start at search at either Amazon or a search engine like Google. With product finding at Amazon, it started it search fulfilled by Amazon. Product finding at Google started at search fulfilled by others. Discovery is emerging as a way to find products, especially on places like Facebook. Instagram started at personalized discovery in the feed and it’s getting more data driven personalized and a lot more competitive. Google in effect is evolving from an ad platform to a commerce platform. Amazon is evolving from a commerce platform to an ad platform. If we look at that eCommerce related advertising revenue on Google, Amazon and Facebook, it is significant four billion on Amazon in the last year of 42% and Facebook has more than 80 million small to medium sized businesses on its site, 23% year to year on its service endeavoring to reach out to customers.
Social media is enabling more efficient product discovery in commerce. A material portion of people that have used social media have found products on social media. A material portion have purchased those products after finding them on social media. If we look at eCommerce referrals from social media, they’re at 6% versus 2% in 2015; rising very quickly. There are large numbers of companies that are emerging DTC retailers and brands that have used social media to get to have very experienced rapid levels of revenue. This looks at the number of companies that have gotten to a 100 million in revenue in less than six years.
With social media, ad engagement is rising represented by Facebook’s eCommerce click through rates which are rising. Return on ad spin, the cost is rising faster than the reach, but both are still rising. The key thing that people are focused on in this area is customer lifetime value. The importance is rising as customer acquisition costs increase. Lifetime value divided by customer acquisition costs as increasely an important metric for retailers and brands. Data driven personalization recommendations are in the early innings at scale. We look at the evolution of commerce over the last number of years, starting with demographic targeting with catalogs, brands, with department stores, malls. Then, we call it utility commerce, transactional commerce on the internet. You search for stuff. Now, it’s increasingly personalized eCommerce, curated product discovery 24/7 recommendations. One of the best examples of that is Stitch Fix.
Product purchases in eCommerce, many are evolving from buying to subscribing. This looks at some of the companies that have subscription service, revenue growth that is pretty and numbers that are pretty compelling. Driven by access selection, price, experience and personalization. One of the best examples of a company in this area is Spotify, which has 45% of its monthly active users are subscribers versus 0% when they launched the subscription product ten years ago. That’s primarily been driven by a really great user experience. Shopping is increasing in the entertainment. Mobile shopping usage is one of the fastest growing areas of app sessions out there. Product and price discovery is often video enabled. Product and price discovery is increasingly social and gamified. Physical retail trending, longterm sales are on a decelerating trend. Then, there’s China and there’s new retail …
Then there’s China, and there’s New Retail, which Alibaba is really leading with. Alibaba is leading in the e-commerce ecosystem, born in China. If we compare Alibaba and Amazon.com, they have similar focused areas up and down the stack. Alibaba has much higher GMV, customer merchandise volume, and Amazon has much higher revenue.
I’m gonna quote some stuff from Alibaba, because not everybody may be familiar with how they view this new area of retail called New Retail, but it’s going to summarize this in some of their words.
“It’s fair to say that our e-commerce platform is fast becoming the leading retail infrastructure of China. Alibaba’s marketplace platforms handle billions of transactions each month, we have the best insights into consumer behavior, we have deep technology, and a comprehensive eco-system of commerce platforms, logistics, and payments.
Alibaba’s increasingly extending its platform beyond China with both acquisitions and equity investments. As of 2007, about 8.4% of its revenue was outside of China, versus 7.9% in the previous year.”
Our friends at Hillhouse Capital have provided us, as they usually do, with some thoughts on China. I’m gonna run through these really quickly.
China in e-commerce, it has number one in e-commerce sales as a percent of total sales, at 20%, highest in the world and the fastest growing. I’m gonna focus on a couple areas of commerce that are unique to China. One is a business called Herbastores, which is basically reimagining the grocery business. I’ll let you look at the pictures for a moment. It looks like a very fun interactive grocery store. The thing that is most interesting about it from a financial standpoint is that a very material portion of their revenue comes from online purchasers where people purchase things online and either pick up or have the item sent to them, and subsequently their transactions per store are materially higher than many of their pure offline competitors.
Also a company called Bell, which is a shoe store, which has censors all over the place. They know where people spend time in the store so they know where to place their products. They have RFID chips in shoes, so they know when people try on the shoes what the percentage take rate is of purchasing the shoes. You can also scan your shoe to get a sense of which shoes may fit you best.
Video and entertainment in China, data usage is rising dramatically, it’s actually accelerating, 162% versus 124% in the previous year.
Mobile entertainment time spent. Mobile video is by far and away the fastest growing area of media and entertainment time spent in China. This looks at short form video, with the blue line and daily time spent in China growing very rapidly. There were seven countries in China that have more than 100 million DAUs that are focused on short form video.
If we look at long form video, the internet enabled budgets surpassed the TV network budgets in 2017. In China there are a number of online long form video providers that have reached tens of millions of paying subscribers in China.
Additionally, an [inaudible 00:15:22] observation is that the number one multi-player video game in China is team based, which is a new thing for the Chinese market over the last five years.
Moving from China, back to the US, which is a little bit of an interesting Segway. Moving it, you want to focus on consumer spending, the dynamics are evolving and the internet is creating a lot of opportunities, in our view.”
For consumers making ends meet, it’s difficult. Household debt is at the highest level ever, and rising. Student debt is the highest grower, auto follows, and mortgage is actually down, versus the previous peak in the third quarter of 2018.
Personal saving rate at 3% versus 12% 50 years ago, and the debt to annual income ratio is rising at 22% versus 15% 50 years ago.
I want to look at how relative household spending has shifted over the past half century. This looks at data from the US Bureau of Labor statistics and compares census data that looks at relative household spending for areas like shelter, taxes, transportation, food. We look at 1970, 1990, and 2017. The things that are rising in a big way include shelter, pensions, insurance, and in healthcare. The things that are falling on a relative basis are foods, entertainment, and apparel.
I want to drill down on food for just a moment, but it’s 12% versus 15% of household spending 28 years ago. Grocery store price growth at the margin is on a declining trend and one of the reasons for that is Wal-Mart. Wal-mart helped produce grocery prices via technology and scale, and subsequently gained a lot of share in that marketplace.
E-commerce is helping reduce prices for consumers. This is a quote from Austin Goolsby at the University of Chicago, “Online prices are falling absolutely and relative to traditional inflation measures like the CPI.”
This is data from Adobe that indicates that over the past town a quarter years, online prices on average have fallen 3% versus 1% for offline prices. The biggest area for Adobe of price declines have been TVs, furniture, computers, and sporting goods. This is a quote from Hal Varied and it says, “We’ve seen how technology can make online shopping more efficient with lower prices, more selection and increased convenience, and this is about to happen to offline shopping.”
Drilling down a little bit more on spending, we indicated that shelter spending has been rising on a relative basis for consumers. That’s what it looks like in the blue, red, and green. Specifically in US, cities are much less densely populated than the rest of the world. South Korea 17 times more densely populated in cities than in the US. Japan 9 times, the UK 6 times. Average home sizes in the United States are materially larger than they are in those places around the world, and US homes are getting bigger while residents per home are falling.
US office space is getting denser and a bit more efficient. Shelter to contain spending, many consumers may aim to increase the utility of space they have. We just provided examples of how some consumers are doing this, airbnb provides income opportunities for hosts. There are an estimated 500,000 individuals in the US that have listings on airbnb, I’m gonna drill into that a little bit later. 5 million listings around the world for consumers. Airbnb can offer lower prices for overnight accommodations.
Looking at transportation, spending is actually relatively flat. There are a lot of reasons for that including oil prices. But consumers are reducing their relative spend on vehicles, and increasing utility with the vehicles that they have. Vehicles stay on the road 12 years versus eight years in 1995. Public transit use is rising and ride share use is rising as well.
Uber is a company that provide work opportunities for driver partners. About a million Uber driver partners in the US, and 3 million around the world and it can also benefit consumers in urban settings, lower commute costs for first person cars for the five largest cities in the US.
Healthcare spending, my partner at Kleiner, Perkins, [Noenoff 00:19:31] pulled this together, I think it’s pretty thoughtful. It takes a look at healthcare spending, and healthcare spending is increasingly shifting to consumers. US healthcare insurance costs are rising for all. US consumers are paying a higher portion of their insurance costs at 18% versus 14% in 1999. Deductible costs are rising a lot as well. The number of employees that have a $2,000 deductible or greater is at 22% versus 7% in 2009.
When customers start spending more they tend to pay more attention to value and prices, especially with things like the internet. Our question is, “Will market forces finally come to healthcare and drive prices lower for consumers?” I’m gonna run through some examples of where it is happening and the trends numbers are still small, but they’re trending in the right direction.
Modern retail experience, One Medical growing its office locations, digital healthcare management Oscar, on demand pharmacy capsule. Women’s healthcare specific solutions, Nurex, transparent pricing, Doctor Consulta, not a US based company, and simplified healthcare with [Cedar 00:20:34]. And we ask the question, “With consumerization of healthcare and rising data availability, may we finally be at the cusp of reducing consumer healthcare spending?” I certainly hope so.
Work is changing rapidly, the internet is helping so far. Technology disruption is not new and technology disruption is adopting the internet. It was adopted faster than PC, faster than the TV, faster than the telephone. What are the drivers of this? Rising and cheaper compute power and storage capacity, and rising and cheaper connectivity and data sharing.
New technologies have created and displaced jobs historically. Concerns that have ebbed and flowed over time. Aircraft jobs have replaced locomotive jobs, services jobs replace agricultural jobs, and agriculture is at less than 2% of the jobs in the United States versus 41% in 1900.
Over the past 70 years, which is the period for which we have data, new technology concerns have ebbed and flowed, GDP has risen, and unemployment has ranged between 3 and 10%.
Will technology impact jobs differently this time? Perhaps, but it would be inconsistent with history, as new jobs and services plus efficiencies, plus per growth typically are created around new technologies.
The job market is solid based on traditional high level metrics in the US.unemployment is at 3.9%, well below the 5.8% 70 years ago. Consumer confidence is high and rising, the index is at 100 versus 87, the 55 year average.
Job openings, at a 17 year high at 7 million, three times higher versus the 2009 trough, and job growth is much stronger in urban areas, where 86% of Americans live, and rising.
Labor course for gestation is at 63%, 3.5 million people below the 64%, 50 year average.
What’s the most common activity for people that don’t work? Leisure, household activities, and education.
Job expectations that are evolving, the most desired non-monetary benefit for workers is flexibility. Technology makes freelance work and other forms of work easier to find. Freelance work is growing three times faster than the growth of the total workforce.
On demand jobs, these are big numbers and the growth is high. We’ve spent a lot of time pulling this stuff together, and this is one of the first areas where we’ve seen it done in this way. On demand workers, 5.4 million estimated 2017 … a 23% year on year estimated to be around 6.8 million in 2018. These are big numbers.
On demand jobs, they’re a bit more than 15 million applicants on the checker background check platform since 2014 for on demand jobs.
This looks at some of the numbers for some of the platforms, we’ve already looked at a few of these. Etsy has 2 million sellers, Upwork has 16 million freelancers, and there are a lot of similar numbers from other players in the marketplace. On demand jobs are filling needs for workers who want extra income, flexibility, and have underutilized skills and assets.
This is data from Intuit that compiles what the general view is on what the basics and the benefits are of on demand work. Number one and two are extra income and flexibility. I’m gonna drill through some of the specifics for some of the companies. At Uber, 87% use Uber to set their own hours, 85% do it for work/life balance.
Their own hours, 85% do it for work-life balance, and 74% drive on Uber to maintain steady income. On Etsy, 99.9% of US counties have Etsy sellers. 97% operate at home. 28% operate from a rural location. 27% have children at home, and 68% say their motivation is the creativity related to Etsy creating and selling provides happiness. Airbnb, 57% of hosts use the earnings to stay in the home that they are listing.
Bill Gurley had a great quote about a month ago that described Uber and the on demand marketplace which I’ll read. “No Uber driver partner is ever told where or when to work. This is quite remarkable, an entire global network miraculously level-loads on its own. Driver partners unilaterally decide when they want to work and where they want to work. The flip side is also true. They have unlimited freedom to choose when they do not want to work. The Uber network is able to elegantly match supply and demand without schedules and shifts. That worker autonomy of both time and place simply does not exist in other industries.” The bottom line is, on demand and internet related jobs, the scale is becoming significant in the US and in other parts of the world.
Data gathering and optimization, it’s been years in the making. It’s increasingly global and competitive. It accelerates with computer adoption, really started with the mainframe in the early 1950s and started with government mainframe deployment. Gathering data for social security, for NASA, for the IRS. Some of the great companies of the last several generations in the US did the same thing, whether it was Bank of America to process checks or Aetna to optimize insurance policies, Visa to create and manage the merchant network or Walmart to track inventory and logistics.
Data gathering, optimization, and sharing is again accelerating with computer adoption. This time, it’s with consumer mobiles and the cloud. We’ve lived through two computing big bangs. The cloud started in 2006 by Amazon AWS and consumer mobile with the Apple iPhone in 2007. This looks at the data behind those platforms. AWS, number of services on the platform and the number of apps in the Apple App Store. Computing big bang volume effects, cloud compute cost declines continue and cloud service revenue is actually accelerating, up 58% year on year.
Data gathering, sharing, optimization enabled by consumer mobile adoption, social media adoption, and sensor pervasiveness. The amount of this that exists is growing at a torrid pace. Data can be an important driver of customer satisfaction. If we look at the US companies that carry a market cap in excess of $100 billion, they have relatively high customer satisfaction scores according to ASCI. The market average is 77 for the fourth quarter of last year, Amazon’s at 85, Google Alphabet’s at 82, Netflix is at 79, Booking.com, Priceline is at 78, Facebook Instagram is at 72.
Google personalization via queries drive engagement and customer satisfaction. Spotify personalization preferences drive engagement and customer satisfaction. A Toutiao … Help me out here. Toutiao … Oh my god, I can’t believe it did that. I am so sorry, but I think I got it right. Interests are driving engagement and customer satisfaction, growing very rapidly with artificial intelligence.
Data improves predictive ability of many services. Data volume is foundational to algorithmic refinement in AI performance. It’s foundational to tool and product improvement, artificial intelligence predictability and capability. Artificial intelligence service or platforms are merging from internet leaders. Amazon.com and Google are increasingly competing in this space. Amazon’s AI platform emerging from AWS enabling easier data processing and collection for others, Google doing the same thing from the Google cloud.
AI and enterprises is a small area but it’s one of the rapidest areas of spending growth in the market. Sundar at Google said, “AI is one of the most important things humanity is working on. It’s more profound than electricity or fire. We have learned to harness fire for the benefits of humanity, but we had to overcome its downsides too. AI is really important, but we have to be concerned about it.”
Data sharing creates multifaceted challenges. Love this comic, but it’s very serious. Data and consumers have a love-hate relationship. The quote is, “Just because I hate you doesn’t mean I don’t love you.” Data and consumers. Most online consumers are willing to share data for benefits. 79% willing to share personal data for a clear personal benefit. Consumers also take actions when the benefits of data are not clear, 64% according to Deloitte deleted certain apps because of data concerns.
As was mentioned last night, the internet companies are making consumer privacy tools much more accessible, looking at Facebook’s change and Google’s change, bringing the privacy tools to the fore from the rear. Data sharing, there are a lot of different views. The EU, Asia, and Americas are rising their regulatory focus on data collection and sharing, while China is encouraging data collection. Cyber threats are also another area where people have a different view on data and privacy, malware volume rising dramatically.
I’m going to spend a few minutes on global internet leadership, the US and China, economic leadership. China and India and the US are the only markets that have relative GDP that is rising, others are falling. Cross boarder trade continues to be very important. Internet leadership, a lot’s happened over the last five to 10 years. This looks at today’s top 20 internet leaders. Five years ago, nine of them were in the US, two were in China. Today, 11 are in the US, nine are in China. China obviously gaining very rapidly.
Smart phones, China is the number one worldwide OEM of smart phones at 40% versus 0% share 10 years ago, while the US has gone to 15% share from 3%. Internet globally, the US platforms lead in numbers, more than two billion users of Facebook platforms and Google, while Tencent and Alibaba have one billion and 700 million, respectively. If we look at those same internet users and look at them by country, the global player that has the largest number of users in each platform is Tencent in China and Alibaba. No one has a peer in the volume of usage and users they have in a single country, other than Tencent and Alibaba.
Feature and data rich platforms, Tencent and Alibaba look at the UI. China internet users are more willing to share data for benefits versus other countries in the world by a wide margin. China’s digital volume is at a significant scale, it’s growing fast and providing fuel for rapid artificial intelligence advancements. If we look at AI in the US and China and look at the competitions, China is increasingly winning the complex tasks in the competitions that take place.
This looks at graduates from bachelors equivalent degrees and doctoral degrees, China in the red, US in the blue, and EU in the yellow, and China trending very strongly. The Chinese government is very focused on developing artificial intelligence. This is a quote from Eric Schmidt, and Cara, you’ll be happy I’m almost done here. “I’m assuming that the US’s lead in artificial intelligence will continue over the next five years and that China will catch up extremely quickly.”
I’m going to close on some economic growth drivers, how they evolve over time. We’re clearly in the period of compute power and human potential. Lifelong learning is crucial in the evolving work environment, the tools are getting better and more accessible. Coursera has 33 million learners, up 30% year on year. The top courses, machine learning, neural networks, and deep learning, introduction to mathematical thinking, algorithms, neural networks, etc.
Lifelong learning educational content usage is ramping very fast. YouTube has more than a billion views of daily learning videos. 70% of users use the platform to help solve work, school, or hobby problems. Lifelong learning, employee retraining is high at some companies. In my view, AT&T is the best example. 77% of its workforce is actively engaged in retraining, most of it web-based. That’s 194,000 people at At&T.
Lastly, lifelong learning, greater than 50% of freelancers updated their skills within the past six months compared with 30% for non-freelancers. I’m going to close with three words. Change, opportunity, and responsibility. We’re living in a period of unprecedented change and unprecedented opportunity. Especially for the people in this room, unprecedented need for responsibility. With that, thank you. Thank you.