• Skip to main content
  • Skip to search
  • Skip to footer
Cadence Home
  • This search text may be transcribed, used, stored, or accessed by our third-party service providers per our Cookie Policy and Privacy Policy.

  1. Blogs
  2. Breakfast Bytes
  3. Mary Meeker: Security, Gigs, Healthcare, China
Paul McLellan
Paul McLellan

Community Member

Blog Activity
Options
  • Subscribe by email
  • More
  • Cancel
mary meeker
Internet

Mary Meeker: Security, Gigs, Healthcare, China

8 Aug 2019 • 8 minute read

 breakfast bytes logoThis is the third (of three) posts about Mary Meeker's 2019 report on Internet Trends. The first was on Monday. The second was yesterday. If you don't know who Mary Meeker is or what this report is, then start at the beginning.

Cyber Attacks

I write about security quite a bit since it affects a lot of technology. It is increasingly accepted that security has to start with a hardware root of trust, meaning on a chip. But Mary covered the big picture:

  • State-sponsored actor attacks are rising, with the US, UK, Netherlands, and Germany all publicly indicting state actors.
  • Large-scale data provider attacks are rising.
  • Monetary extortion cases are rising, partially enabled by cryptocurrencies enabling anonymous ransom payments. Most high visibility were the British National Health Service and the city of Baltimore.
  • Attack-to-detection dwell times continue to fall, a good thing. In 2011 it was 416 days (over a year). In 2018 that was down to 78 days (2.5 months, still way too long).

The state of security is still really low, even in financial institutions. For example, FireEye Threat Research was engaged to help evaluate security and response capabilities at a financial services company and reported in April:

The team was able to obtain full administrative control of the company domain and compromise all critical business applications without any software or operating system exploits. Instead, the team focused on identifying system misconfigurations, conducting social engineering attacks and using the client’s internal tools and documentation.

The result of this sort of thing is that 31% of datacenter operators experienced downtime in the last year, with a big increase in the number of sensitive records exposed:

 Unemployment and the Gig Economy

When asked if unemployment and jobs were among the top three worries for their country, the responses ranged from 69% for Italy down to 11% in the US (which is 3 times the unemployment rate of 3.8%, a 19-year low, although the way that is calculated excludes people who are underemployed).

.

 One aspect of unemployment where the internet is creating opportunities is on-demand work (aka the gig economy). On-demand consumers doubled in the last two years in the US to 56M people. This includes food delivery (e.g., Doordash, Uber Eats, housing (Airbnb), online marketplaces (eBay, Amazon), transportation (Uber, Lyft), and more. This has driven on-demand platform workers up from 2.4M in 2015 to 6.6M today. Checkr, who helps employers find on-demand workers, has handled over 20M applicants in the last four years.

About half of people doing on-demand work are motivated by being unemployed, but there are many other reasons:

 The growth rates of these on-demand platforms and internet-enabled marketplaces are huge: Uber has 4M drivers, Etsy has 2M sellers, Airbnb has 6M listings.

Online Education

One area I've been surprised hasn't taken off more is using the internet for education. I think there are probably two main reasons. One is that you have to be very self-motivated to just educate yourself using internet resources (or even just using good textbooks in earlier eras). The other is the certification problem, that you largely get a degree for simply attending a prestigious university, not for passing an external standard analogous to the bar exam. In fact, in most states, you are not allowed to take the bar exam even if you can pass it without attending law school, even if most of what they teach has nothing to do with the bar exam. There is a debate about what is called the signaling model of education, about whether what you learn at university is actually valuable, or whether just having the degree certificate from a prestigious university is what is important, whether you learned anything or not. If you study engineering, there's a good chance that you will use a lot of what you learned in your career. If you studied history, probably not.

The trend is clearly for more and more institutions to have significant online offerings, including several prestigious universities.

One of the leaders in bringing these courses online is Coursera. The numbers are all up and to the right, closing in on 40M registered students (or “learners,” as Coursera likes to call them). They are roughly evenly split between North America, Asia, and the rest of the world (Europe, South America, Africa). Udemy has another 32M students (many of whom, of course, might be on both Coursera and Udemy).

Fast-growing, and getting lots of news, is Lambda School. They have a unique business model in that you pay no tuition and only pay for your education in arrears once you get a job paying $50K or more. They are still small... but on the other hand, they only started about 18 months ago.

It is not specifically educational, of course, but one major channel is YouTube. They have 4.5B (billion with a B) hours of how-to video content. Over 200M views on Booktuber content. 59% of Gen Z users say YouTube is their preferred learning tool. Some educational names use YouTube as their vehicle. Khan Academy has 5M subscribers, TED-Ed and Crash Course both have 9.5M subscribers (I updated the numbers from Mary's presentation with what YouTube displays today).

The challenge with a lot of online education, as I mentioned at the start, is self-motivation. This shows up in completion rates. These massive open online courses (MOOCs) have variable completion rates (Coursera numbers):

  • Unpaid learners just 4% finish non-degree consumer courses
  • Paid learners 50% finish non-degree consumer courses
  • Enterprise learners 44% finish enterprise courses
  • Paid degree learners 89% finish degree courses

 Digital Healthcare

Healthcare is a big expenditure in many countries. In the US it is rising fast, both in the federal government and the states. The graphic shows the percentage of GDP spend on healthcare, with the grey bar showing government/compulsory health expenditure, and the blue bar voluntary.

In the US, there is strong adoption of digital health tools. About 30% of the US has adopted wearables. Genomic testing has grown from basically nothing in 2012 to about 10M today. This is also proving a growing base of genetic data. Usage of electronic health records (EHR) has grown to close to 100% in both hospitals and physician's offices. Mary's chart shows the top physician "wants" are interoperability, predictive analytics, and reduced cost. But anecdotally I hear that almost all physicians hate EHR since they get between them and their patients and consume far too much of their time.

There is a lot of low-hanging fruit in healthcare. Just scheduling patients effectively, for example. Patients tend to wait over 3 weeks to see a doctor, and online scheduling can reduce that to 24-72 hours with companies like Zocdoc and Solv.

Telehealth, providing remote medical care, is another use of technology. Last year (globally) there were nearly 3M Teladoc visits, for example.

Doctors can use technology among themselves too, with Doximity providing a physician social network, now with 1.2M registered members up from 0 when they started in 2011. For some rare diseases (getting bitten by a Macaque monkey and getting encephalitis, for example) there may be as few as 20 specialists in the whole US who are experts. If you are a normal family practitioner you almost certainly don't know any of them personally, or even have one in your area.

But I think that the biggest change is perhaps that the internet giants are both trusted with health data, and are providing various aspects of care:

  • Google is working on AI-enabled healthcare and even devices like contact lenses that can monitor insulin.
  • Microsoft is also working on machine learning for healthcare.
  • Amazon/Pillpack are working on consumables like drugs and inhalers. It turns out that in the US, 62% of people take prescription drugs, and 24% of people take over four.

China, Home of the Super App

As in previous years, Hillhouse Capital prepares a section for Mary on China Macro Trends. I'll just pull out a couple of slides that caught my eye.

Urbanization. China has gone from a few percent living in cities back in 1950 to over 50% today. That started as less than India and is now almost twice India's percentage. But there is a lot of room to grow to reach US and Korea, let alone the most urban country of them all, Japan at around 90%.

 There is very strong China internet usage growth driven primarily by short-form video (and almost all China internet access is mobile). Note that in these charts, data usage goes back to 2012, but the short-form video graph is just over two years:

 One big difference between US/Europe and China is the way platforms, such as WeChat, Alipay, and Meituan, provide everything in one place from chatting, to browsing things to buy, to placing an order, to payment. All without leaving WeChat.

You can think of Meituan as one App that is Yelp, OpenTable, Fandango, Airbnb, Hotels.com, grubHub, Square, Kayak, and Whole Foods all rolled into one. Meituan started as a Groupon clone, and now provides over 30 services to over 400M users, growing at 26% per year.

Alipay is another superapp, with over 200,000 min-programs, over 1B users, and nearly 300M daily active users.

Learn More

There is a huge amount in this presentation that I didn't even touch on, and even in areas where I pulled a few significant details out, there is a lot more. Here's the full 333 slide deck.

 

Sign up for Sunday Brunch, the weekly Breakfast Bytes email.