72 comments

[ 2.6 ms ] story [ 137 ms ] thread
I have to vehemently disagree with slide 36: "It's Crucial to Manage For Unintended Consequences . . . But It's Irresponsible to Stop Innovation + Progress"

I'm afraid that's exactly backwards. It's Irresponsible not to manage for unintended consequences... but it's crucial to keep innovation and progress moving forward.

American capitalism vs European capitalism at its best.
Can you eloborate on that? Things like “The Lean Startup” (de-risking as principle) originated in the States
Irresponsible implies a moral imperative. I think they're saying that "unintended consequences" often harm other people, and therefore business have the responsibility to minimize them. The stereotype is that American business like Uber, Facebook, and Amazon tend to see shareholder value as their one and only duty.
I’ve always been very skeptical at this report. I’m not exactly sure why, but I think because a) it is mainly a collection of a lots of data and b) it is extremely popular with many business people who often have little knowledge on the subject
I'm also skeptical of reports where one seems to need 294 slides to identify a trend. But drowning viewers in data and having no conclusions at the end is all the data I need.
True. What I don’t get though: Why is it sooo popular then?
It's popular because it's a convenient compendium of a bunch of charts.

Let's say I'm wondering about "number of smartphone users growth". A google search might pull one chart for that.[1]

What Mary Meeker does is ask 200 of those types of questions, gathers all the data to make the charts, and puts them together in one deck. That's more convenient than conducting 200 searches on the Statista website.

I think the disconnect and dissatisfaction happens because they expect Mary Meeker's slides to provide some mind-blowing insight. I think it's more realistic just to accept it as a nice data dump of some of the latest trends.

[1] https://www.statista.com/statistics/330695/number-of-smartph...

Thank you for this explanation. Makes completely sense
Not expecting mind blowing insight, just insight. Data turned into knowledge and models, showing relationships and significance. Gathering data is an intermediate step in this process. But true, it is a nice dump of data!
It helps to have an alliterative name. ;-) Only half joking. The executive types who like this kind of thing also have a tendency to think alliteration makes them sound smarter.
Mary Meeker became popular about seven or eight years ago, and at the time, there was an executive summary: "mobile is much, much bigger than you realize."

At the time, many people didn't realize that Android + iOS shipments had dwarfed Wintel shipments until they saw this dramatic image in her 2012 slide deck. https://image.slidesharecdn.com/internettrendsstanford120312...

Since then, IMO, people keep showing up because they're looking for another major shock. But there hasn't been a major shock like the rise of mobile since then. Mobile is still big. China is still big. etc.

Now her presentations convey a feeling of comfort. "I believed all of this already, because I am a smart, well-informed person, and here's the data to prove it."

Mary Meeker became popular in 1995. She was the lead internet equity research analyst at Morgan Stanley and covered the IPO of Netscape. She was one of the top equity research analysts in 1998-2000 covering EBay, Amazon, Drugstore.com amongst others
I think the alliteration in her name and the fact that it's an annual report ("Oh yeah, I read this last year") contribute too.
Speaking for myself: Because it's a collection of lots or relevant data in one place. You get to pop up your head and look at the world at large. How is the industry as a whole progressing, what particular areas are "hot" in terms of attention.

It's well researched, it's sourced, it's free - quite likely the only compendium of that size that meets those criteria.

>need 294 slides to identify a trend.

Her report isn't trying to identify "a trend" (singular). They are slides for multiple trends (plural). Notice that "trends" is the actual title of the report.

There isn't meant to be only one single narrative for all the slides. (Well, the "internet" is a common denominator but "internet" is obviously already in the title.)

E.g. the slides for voice assistants trends (Amazon echo) are separate from housing space utilization trends (AirBnb). Multiple trends for different technologies, different business sectors, different geographies, etc.

(comment deleted)
The report has value to find trends or insights that may apply to a particular world, specifically when trying to get a sense of disruption.
Also the y-axis on a lot of the charts seem arbitrary -- designed to show spikes to underscore assumptions.
That, and the fact that it comes in the form of a slidedeck makes it hard to do itemized critiques. I guess it wouldn't be hard, if someone cared enough, to manually go through the slides each year and note them in a spreadsheet. But the other problem with slides is that it is very cumbersome to find the source data used for the charts. Hell, it's hard to even read the footnotes for these charts, even for this relatively simple one:

https://www.slideshare.net/kleinerperkins/internet-trends-re...

edit: Slide 8, "Internet Penetration"

As an example of why the insights seem kind of bland to me, here's where she has a pie chart about "Where Do You Begin Your Product Search?", with Amazon at 49% vs 36% for "Search Engine":

https://youtu.be/HdjcdZqODoE?t=8m22s

However, looking up the source for that data, and it seems like the more interesting insight is that Amazon's lead shrunk from 55% to 49%:

https://www.bloomberg.com/news/articles/2017-12-20/google-na...

But I guess for most people, it's insightful enough to know that Amazon has such a lead in the first place? I'll admit I didn't know that without seeing her slide.

> the fact that it comes in the form of a slidedeck makes it hard to do itemized critiques

The slides usually cite sources, so it's pretty easy to tell if the underlying methodology is weak. (And often it is.)

I usually enjoy reading the annual reports, my biggest criticism is just that it focuses on completely different trends each year. It will go from social to AI to messaging to privacy, etc. If something is only relevant and worth talking about for one year, then pretty much by definition it isn't a trend.

Heh. Trend is a fun word since it both means "a continuing movement" and also "a fad."

"That's a meaningful trend in the data" vs "it's just a trend."

My criticism is that its usually US -centric, sure there is more about China every year, but the healthcare section is all US, lots of media usage habits which are all US, economics, GDP all US.

For example, I've not learned much at all from that deck about whats happening in Africa or India.

>But I guess for most people, it's insightful enough to know that Amazon has such a lead in the first place? I'll admit I didn't know that without seeing her slide.

I think you nailed it here. It's only intended as a starting point, for drawing your attention to a whole range of data points you may not already been aware of. Getting caught by one slide or data point, wondering what the deeper story is and going off and researching it yourself is exactly what the deck is for.

Normally I would be, but it seems they have access to special data b/c they fund companies like Coinbase and NextDoor.
(comment deleted)
The report from past five years:

Most successful predictions obvious (mobile, China). Lots of misses (AI, cloud computing), and lots that’s wrong (everyone will be using wearables?).

On slide 25, she claims Google Voice recognition has reached 95% accuracy, which seems to be human level threshold.

“Google Machine Learning Word Accuracy”

Does this mean anything practical to users?

Accuracy is important, and it's steadily improving. I now prefer to dictate my comments and essays instead of typing, especially on mobile.
The garbled voice mail translation I still see would beg to differ. Maybe on clean audio sources they have? But they still translate some absolute nonsense that I can easily understand in one listen.
I'm not sure all Google products use the same voice recognition tech. Almost sure that they're not.

In the NLP realm for instance, Google translate is (or used to be) based on a much older statistical model rather than the hot deep learning models the cool projects use nowadays.

Human level? You miss/misunderstand one word in 20? It needs to be much better than that to be human level!
So, what's the context for this for those of us that don't know? Who's Mary Meeker, and why should I care about her trendspotting? Just wondering if it's worth it to wade through 300 slides.
she's a partner at kleiner. if you want to hob-nob with VCs, you should be familiar with this report so you can look like you're in the know.
If only there were a way to search for basic information. Sorry :-)

Mary was one of the early financial analysts to focus on the emerging Internet. She's now at KPMG and this report is an annual distillation of a lot of the current trends and data.

KPCB, but easy mistake
Never make a mistake in a snark :-)

I mess that particular one up all the time.

Mary Meeker was a Managing Director at Morgan Stanley way back when that investment bank did the IPO's of the first internet generation including Netscape and Google ... She has been in the know since before the beginning ;-)
Most of the time, the value I derive from this report is to gain a directional sense of the tech world that can then be combined with other insights to form actionable items.

For e.g. in terms of investing, smartphone shipments are not growing, which means that for companies like Apple that derive a significant piece of their revenue from smartphones will probably have to find growth elsewhere, either further up the stack or by launching newer products. This means that until a significant non-handset source of revenue emerges, Apple would likely not grow dramatically from this huge base.

This report is always interesting, but watch out for Mary's right-wing political views she likes to add in.
A section on the US national debt/spending was the only remotely right-wing thing I saw.
No mention of blockchain or bitcoin or ethereum or cryptocurrencies. Really ?
There's a slide on coinbase membership
Oh boy. That is one tortured slide. "We must have a hockey stick, find me one".
Not enough imo. The space is much more beyond coinbase membership
This was pretty good last year but this year isnt forward looking at all.

Also they missed the 100 mn new internet users which came online in India the past year and the rock bottom mobile data rates which enabled it.

This report contains a lot of data. Is there a counterpart to it that, for each slide, attempts to answer the question "So what?"
"0% new smartphone unit shipment growth in 2017"
That, combined with the mobile market growth. Means people are happy with their phones, find no reasons to upgrade.
There is no relation between growth and renewal. Say there is 3BN mobile users and 3BN shipment a year. That's each user renewing every year. But next year you only have 3BN shipment.

-> 0% growth + people upgrading every year.

No mention of healthcare at least in the TOC this year, perhaps not surprising given Beth seidenbergs recent departure. Anyone know if they're still doing HC stuff?
There is definitely mention of health care, including speculation that tech may help bend the cost curve down.
Look at her report from five years ago before hyping her predictions. Most successful predictions obvious (mobile, China). Lots of misses (AI, cloud computing), and lots that’s wrong (everyone will be using wearables?).
If you count Apple Watch and various fitbit items a lot of people are.
No. iOS in general has low market penetration, let alone wearable.
Maybe your very inner circle. Hop on in the Metro and see how many are wearing Apple watches.
Apple watches I agree have a very low market penetration, but Fitbits and related products seem to be quite popular, although based on anecdotal observation their popularity does seem to be down a bit from their peak.
Insane growth of all the messenger apps surprised me.
Pretty light on in the enterprise computing section. I had expected more, enterprise applications are far more than just dropbox/slack/some product I never heard of before this report.