Uber The Colonist

It seems that Silicon Valley is at last waking up to what Uber really is.

  1. “Monopoly as the Uber Business Model”
  2. “Understanding That Unregulated Monopoly Was Always Uber’s Central Objective”

I am a bit disappointed that it took this long for Silicon Valley to see this, but I suppose better late than never.

Over a year ago, I wrote this;

The question is will Uber be a sustainable business? Will it raise prices after venture capital runs out and there is no competition left? If they are forced to employ their drivers as employees and if they have to also pay for their driver’s cars, which is quite possible long term, can they still maintain current prices? If Uber becomes a monopoly, will they be any better than the regulated monopolies before them for both the drivers and the customers? I have serious doubts on this, and unless Uber discloses the sustainability of its business, commits to future low prices and the welfare of its drivers, I think that strictly regulating Uber makes a lot of sense. The last thing that you want is for Uber to kill your local taxi industry, and replace it with one which is just as expensive (potentially more) and where all the profits are funnelled to a Silicon Valley company far away. This is why we have anti-trust laws, for example, and this is why we regulate industries (like the public transport, mail, health and food industries) that directly affect the welfare of our citizens.

The point that I want to emphasise is that if the US is killing itself as a result of its relaxed views on anti-trust and disdain for regulation, then so be it. I do not mind the world’s largest superpower shooting itself in the foot.

I am however not OK with how the US is exporting this to other countries. If Uber is killing local taxi industries in developing nations, preventing the deployment of public transport by providing an artificially cheap option, and in general making these countries dependent on the US for basic needs, then I see this as a new form of colonialism. This is what Gandhi fought against with the Swadeshi movement.

And we should also note that this is not restricted to Uber. One could argue that the stagnation of tech in developed countries has caused Silicon Valley giants to search for growth in the developing nations, and their huge resources are allowing them to use predatory, money-losing tactics. It’s just that since the US is inherently an inwards-looking country and nowhere near being truly cosmopolitan, they don’t realise how much damage they’re causing.

Just see how much China’s Internet has prospered by shutting out Silicon Valley.

If Silicon Valley wants to earn money in developing nations, I see no problem in doing so. However, they must compete on equal terms. They must earn profits. For example Apple is OK because even in developing countries, they charge the same price (which turns out to be super-premium in these places). Apple does not drive out local competitors, but encourages them to copy and provide the same features at lower prices (again, look at China). Local cheap competitors thrive because of Apple.

Predictions For 2017: iPad Sales Growth

This is the second in my series of posts where I make predictions for 2017. The first one was about Autonomous Driving.

iPad sales growth

2016 was the year when we started to see revenue growth (but not unit growth) in the iPad. Many were quick to say that this was due to the introduction of the iPad Pro, but I think this misses the fundamental dynamic of what is happening in the tablet market. In fact, I have said in this blog multiple times, that most tech pundits have not understood the dynamic of the tablet market from the very beginning. The people who attribute revenue growth squarely on the iPad Pro inevitably expect a very slow growth going forward, since they do not see continuous growth drivers. My prediction is different in that I expect accelerated growth that will be in the high single digits.

Here I will illustrate my thesis and show why we should expect strong growth in 2017.

名称未設定 numbers

The above chart shows my hypothesis for what has been happening in the iPad market from the beginning; why we saw a very strong introduction, followed by a decline, and then a plateau.

  1. First of all, I separate the iPad market into two distinct segments. The first is the “Entertainment” segment which includes gaming, video watching, etc. The second is “Productivity” which includes writing, drawing, video/audio production, etc.
  2. In the initial phase of the market, we saw a huge uptake of usage in the first “Entertainment” segment. Even though the iPad was a new category device, looking at its gaming and video capabilities, it was a clear and obvious replacement for mobile game consoles like the Play Station Portable and the Nintendo 3DS. It was also a simple replacement for secondary TV screens. Since consumers could easily see the benefits and how it would work, the initial adoption was very rapid. That is, there was no need for an early adopter phase where only a fraction of the population would understand the merits of the device.
  3. However, as smartphones gained processing power and larger screens, they also started to satisfy the “Entertainment” segment. Hence the later decline in sales for this segment which started to happen in 2013-14.
  4. All this while, the “Productivity” segment of the market was going through a regular adoption curve of new category products. That is in the first few years, only the brave early adopters used iPads for “Productivity”. However, the number of these users has slowly but steadily been rising. In many cases, this has been happening more in the corporate market than in consumer markets because frankly, “productivity” is more important for our work than for leisure. It is important to note that whereas larger screen smartphones are adequate for playing games and watching videos, it is really torture editing a spreadsheet on smartphones. The benefits of a larger screen tend to be more pronounces in the “productivity” segment.
  5. Therefore, looking at the sum of both segments, we will see something like the yellow curve where a period of decline will be followed by steady growth.

Although I have made the “productivity” segment to show linear growth in the above chart, in reality, it is more likely to be sigmoidal. Therefore, when the “productivity” segment gains steam, we are likely to see quite steeper growth.

From my thesis, I can predict the following;

  1. We will see strong growth of the iPad in 2017 onwards. 2017 will start slow, but growth will accelerate.
  2. Since growth will come from “productivity” segments, the seasonality of iPad sales will become less severe.
  3. We will continue to see strong sales coming from corporations, but sales to consumers may continue to be weak.

Since 2017 is still the early phase of “productivity” segment adoption, it might yet be a bit early to see a strong impact in 2017Q1 and Q2. However, I do expect 2017Q3 to show a significant effect. 2017Q4 will be less impressive due to the “entertainment” segment dominating during the holiday season.

Predictions For 2017: Autonomous Driving Reality Check

I am planning a series of posts where I make predictions for 2017. I will put each prediction out one by one, and I will only pick those that have a strong implication for how we think about tech and innovation in general. I will also try to pick those that are likely to actually happen in 2017, rather than something that will happen eventually. That is to say, I will make it possible to check if the prediction was correct at the end of 2017.

Serious autonomous driving fatalities

In 2016, we saw a Tesla owner killing himself in a self-driving car. We also saw Uber self-driving cars running red lights in San Francisco.

Tesla managed to wiggle out of the problem by putting the blame on the driver, who may have been watching a Harry Potter movie instead of being ready to resume control of the vehicle. Uber managed to put the blame on the driver, by saying that the driver was actually in control of the vehicle at that time (which frankly sounds rather unconvincing).

In 2017, more companies will put their self-driving cars into public roads. Fierce competition and investor pressure will mean that some companies will even do this prematurely, before the technology is truly ready. In effect, it is likely that we see something like the Titanic crashing into an iceberg. That is, we will see companies hastily putting autonomous cars onto roads before they are ready, possibly with more fatal consequences. For the sake of prediction, I would say that we will see at least two fatalities by June.

What will subsequently happen is very politic and depends on the huge lobbying power of the large tech companies. There will no doubt be a move towards regulation, but on the opposing end, we will also see an eagerness from governments to embrace the promise of innovation. It is difficult to predict which way the scales will tip.

Alternatives To Amazon Go That Already Work

Amazon recently announced Amazon Go located in Seattle, which is currently in beta phase and which is planned to open to the public in early 2017. It is chock-full of image recognition and AI, suggesting that only the US tech giants investing tons of money in software can implement similar solutions.

Well guess what. Similar stuff is already being worked on in Japan. Announced yesterday, Panasonic and Lawson have announced an automated checkout system that puts the barcode scanners into the shopping baskets so that customers can scan while shopping, instead of waiting to scan at self-checkout registers. They have also announced that they will attach RFID electronic tags to each piece of merchandise starting February 2017, thereby eliminating the need for even the barcode scanning (you just put the items in the basket).

Although the current system still requires you to checkout at a register (which puts your groceries into plastic bags for you while you pay), the time required will be significantly reduced and hence the queues will be shorter.

More significantly, this system does not need large numbers of cameras scanning your every movement and is not creepy. You don’t need a smartphone and you don’t even need to give the shop your identity.

Going forward, it is clear that RFIDs are better suited than barcodes and could provide similar experiences to Amazon Go by having an RFID scanner located at the shop exit. RFIDs are still a bit expensive (I hear they cost about 10 cents), but if their usage scales, then we can expect this to come down significantly.

I hope this clearly demonstrates that AI is not the only solution to the issues we have in life, and there are often other less creepy but equally effective ideas out there. Notice that RFID is not new and that it is already a 10 billion USD market.

Update

Note that for example, the convenience store retail chain 7-Eleven is reported to have beaten Google and Amazon to the first regular commercial drone delivery service. Although I am sure that Google and Amazon are working on a more technically complex and advanced solution, this clearly illustrates that you do not have to be a tech giant to make these things work.

Imagining An Attack That Could Wipe Out Trust In The Cloud

In a previous post, I described how the public confidence in services that collect your private information is getting closer to crossing the creepy line, and that any security breaches that actually harm customers would likely be the final straw that break the camel’s back. I also mentioned that instead of Google, Facebook and others, I predict that the final straw will actually come from hackers intruding into our devices and accounts.

Here, I want to elaborate on the example that I have in mind. Before this however, I want to update the reader on a relatively new form of malware called Ransomware. Ransomware is holds your data hostage and importantly, instead of just causing you trouble, it blackmails you to send money. Also of great significance is that Ransomware creation can now be outsourced.

Ransomware works because victims are willing to pay money to get back their files. However, now that we often have more valuable data on our smartphones or in the Cloud than on our PCs, it is reasonable to assume that hackers are right now thinking of new ways to hold your private photos, your location data, or messages that you might want to keep secret as hostages.

For example, a recent Apple Ransom scam  asked for a $30-$50 ransom or otherwise they would do a factory reset. The author advises that you simply ignore this because you can easily recover with a backup. However, what if the scammer had threatened to publish all your photos, your emails, your location data, etc. on the web for all to see. Would you still ignore the scammer? Unlike the iCloud celebrity photo leak, this is something that could happen to any normal person, and this is what makes it so scary.

This is not about Apple vs. Google/Facebook or about any single company’s approach to privacy. If such attacks became widespread, it could cause people to be scared of storing anything in the cloud, despite whatever security measures each individual company took. Of course two-factor authentication will help, but not enough people use it yet.

Advanced two-factor authentication systems may mitigate the worries in the future. However, if such attacks strike now, I fear that the companies that depend on the cloud will have a hard time getting people to trust them once again. Given the potentially widespread consequences, I think this is definitely something to give due thought to.

The Last Straw Of Creepiness

Eric Schmidt has previously mentioned that  Google’s company policy was “to get right up to the creepy line and not cross it”. And despite occasions where it has got itself real trouble by sniffing in on data from personal WiFi routers as Google Maps cars roamed the streets, the general public has not made a huge fuss over privacy.

However, it is not as if the general public is totally oblivious to the privacy issues. There have been reports, for example, that young adults (being more tech savvy) take more security/privacy measures than their elders. Interestingly, the young adults are more concerned with hiding information from their parents than from Google and Facebook, which is obvious when you think about it. A Pew Research survey also show “Some 86% of internet users have taken steps online to remove or mask their digital footprints, but many say they would like to do more or are unaware of tools they could use.” I have also seen similar surveys in Japan.

If this is the case, then it seems like there is a delicate balance in place which reflects Eric Schmidt’s quote, where online privacy is a serious issue for many, but not quite enough for a public backlash. The Internet trackers are a whole have been successful in coming close to the creepy line, but also in not having crossed it.

The next question is then, how will the Internet trackers including Google, Facebook, and a slew of other online ad brokers, cross the line? When will they do something that is so creepy that the public will revolt?

The key to understanding this is, I think, by recalling why the US does not appreciate Edward Snowden. Snowden uncovered rampant privacy intrusions on a massive scale by the NSA. However, the US citizens do not seem to care so much. They seem happy to let the NSA collect data, as long as the information spied upon is not used against themselves, but against terrorists. Of course, I’m sure that US citizens who come from the middle east are not so reassured, but for the majority of Americans, they simply don’t consider themselves as the victims.

Looking at it this way, the creepy line will be crossed when and only when the massive data collected is used against the majority of citizens, and not against terrorists, in a way that is easily noticeable and potentially harmful. For example, re-targeting ads are getting very close to the line because they demonstrate in an unambiguous way, that Google is carefully watching which 3rd party websites you visit. This is completely unlike the previous generation of search or display ads. Re-targeting ads have reminded the public that Google is watching your every move. The only thing that has to happen now is for something to demonstrate that this information can be used to harm you, and then the creepy line will most likely be crossed.

Thus we should next focus on when the public will consider the information gathered by Google and Facebook to be dangerous and harmful. If the information that they have is used in crime in a way that the majority of citizens can identify with, then I would most certainly expect a backlash. This will be when the creepy line will be crossed. However, Google and Facebook themselves have no intention of harming their users, so it won’t be them that cross the line. It will be someone else.

There is no doubt that the information in Google’s servers is potentially damaging. Google probably has the most harmful data if revealed. Unlike Facebook or Apple where you typically send the information yourself, and are unlikely to send stuff that will harm you down the road, Google collects everything. They collect all your searches, all the places that you’ve been to, and all your emails. You do not select which information to share with Google, so the good and the bad get sent there.

I think we a just one major security breach or one major malware attack away from a crisis of confidence. Google itself will not cross the line, but malware can make this happen. Current malware does not collect the privacy/location information from Android devices or Google accounts, but this is because the business model is not there yet. If somebody decides that this is indeed something that they can make money from, then this will happen, and I expect it will bring down Google’s data collection practices down along with it.

Security breaches are becoming more sophisticated and more targeted. Large leaks of accounts are reported quite frequently, although not all of them can be entirely trusted. Indeed, one could imagine hackers announcing the leak of a large number of bogus accounts, just to scare the public into responding to phishing emails. As long as this trend continues, I believe that the largest threat to Google’s data collection practices will be a security breach and not a sudden awakening to privacy by the public.

Points Of Convergence

I’ve been studying the Android operating system and ecosystem recently, out of boredom, and one thing struck me as very odd. That is, Google’s AI applies (or is supposed to apply) to Google’s services, but that is not necessarily the case for third parties.

For example, the Calendar app has an option to scan Gmail and extract event information (presumably using clever AI), which is then automatically added into your calendar. However, although the Android Gmail app can now comfortably connect with Microsoft Exchange servers, Android cannot touch email coming out from these.

This is not the case with iOS. Mail.app analyses all email, regardless of where it came from, applies the same algorithms to extract date information, and suggests events to add to the calendar (it will automatically add events if there is an attachment with the “.dat” extension). Mail.app can do this legitimately because everything is done on the device, without sending the emails to Apple servers. (Note that Google is still facing litigation regarding scanning of email for advertising purposes, which questions their scanning non-Gmail originating email.)

From a technical point of view, features like Google Now-on-Tap may allow Google to analyse data that resides in third party services. However, these third parties may be reluctant due to competitive concerns. Furthermore, privacy policies at least in Japan are very sensitive about sending data to other entities. I expect the same policies or at least expectations exist in many other countries as well, and the above Gmail litigation suggests that this is indeed the case.

This means that iOS will be able to analyse and learn (locally) from emails stored in third party servers like Microsoft Exchange, whereas Gmail will not. In other words, the iOS Mail.app can stand at the crossroads where information from multiple sources come together, and learn from each of them. The iOS Mail.app will be able to benefit from being at the point of convergence. On the other hand, Gmail will have to be separated and isolated.

Gartner has reported that only 4.7% of public companies use Gmail for work. This is based on email routing records, and is likely to be quite reliable. Therefore, in terms of the treasure trove of corporate email data, Gmail is mostly insignificant. For Google to really access this information, it most likely needs to move away from cloud computing and towards AI on the device, because the device is where the data converges.

The interesting to note here is that the point of convergence from a purely technical point of view, will not necessarily be where it will be in real life. Whereas technically it makes sense to accumulate all data in the cloud, privacy concerns alone could force convergence to occur solely on the device.

What Is AI Good For?

With all the hype surrounding artificial intelligence, it is important not to get too immersed in the technical and science fiction aspects. Steve Jobs said “You‘ve got to start with the customer experience and work backwards to the technology.” and this is true of artificial intelligence as well. Although a full discussion is totally out of the scope of a short blog post, I would like to provide a few perspectives.

Contextual interfaces

In my opinion, the right click in Windows 95 was a huge innovation. Prior to that, one had to look inside a huge array of options under a menu bar, or scan through a panel of small and often obfuscated icons. The right-click contextual menu showed a short list of tasks that were all relevant to the object that was currently selected and relieved you of this wasteful routine.

Although this may not strictly be classified as AI, the way that that it lessened the burden on the human brain was significant.

Similarly, one application of AI that would most certainly be very popular with users would be UI improvements that significantly reduced the need to scan through a list of options to find the relevant actions. In iOS 10, Apple has introduced AI that learns which emails should be sorted to which folders and intelligently provides a shortcut so that sorting emails is much quicker and easier.

Automated processing

Email spam filters learn what emails have a high probability of being span. From a user perspective, this is by all means artificial intelligence.

Although spam filters occasionally make mistakes, they help save our time and cognitive load by pre-filtering out stuff that is completely irrelevant to our work. Good spam filters also protect us from phishing attacks which can compromise whole corporate networks, and so it is no surprise that these are in high demand.

This is a very important market for AI.

Data Detectors

Apple had a patent for a very powerful technology commonly known as Data Detectors. This technology can detect addresses, event dates etc. inside text, and dramatically improves the user experience on smartphones where it is inconvenient to copy and paste.

The analysis of text, prediction of what a user might want to do with it, and providing a convenient and intuitive UI that enables the user to quickly get it done, can be a great timesaver.

Voice Recognition

It is well know that machine learning techniques have greatly improved voice recognition. Voice recognition has historically been valued by people who have difficulty typing. With mobile devices, voice recognition is convenient when you cannot use your hands.

Voice Interface

Graphical user interfaces are great for a stepwise approach for getting things done. However, since they operate by providing a list of options on a 2D screen, there is a limit to the breadth of commands that can be issued at any one time.

Command line interfaces and voice interfaces can get around this issue because they do not have to present a list of options. They are limited only by the ability of the user to memorise the available commands, and to issue them without referring to a menu. Hence voice interfaces are a convenient way to issue tasks quickly.

Summing up

My intention with this post was to show that there is much more to AI than a voice UI, and that from a practical perspective, the other applications have already proven to be very significant in terms of user benefit. Although voice UIs and predictive assistants like Google Now are interesting and futuristic, there is no reason why these applications will be the most useful and revolutionary.

Current advances in machine learning (Deep Learning) will build upon what we already have, and for smartphones with big screens, what we already have is a good graphical user interface.

AI, Voice UIs and predictive assistants should be evaluated based on their merits. How will they save us time and for what tasks? How will they help us when we cannot or it would be inconvenient to view our phone screens? How can they reduce our cognitive load?

Apple is pretty good at understanding what the user experience should be, and arguably, this will be just as important or maybe even more so than the underlying algorithms.

Peak Google Revisited

Almost a year ago, I noted that while a few prominent tech pundits had pronounced “Peak Google” at the beginning of 2015, Google was actually as strong as ever 12 months later.

In my post, I said that since no company keeps succeeding forever, anybody that predicts the demise of a company without giving a specific timeframe will always eventually be right. That is to say, any prediction without a timeframe is utterly valueless. I also noted that giving a timeline is extremely difficult.

However, I think we now have enough information to give a rough timeline on when we can expect “Peak Google” in financial terms.

Data points

I will lean on the following data points.

  1. The historically constant size of advertising spending
    In 2014, Eric Chemi writing for Bloomberg noted that the US advertising industry has always been about 1 percent of US GDP since the 1920s. This is significant because the US is much wealthier than it was 100 years ago, and it has gone though many ups and downs, even one world war in this time.
  2. The share of internet advertising within the whole advertising market
    According to eMarketer, total digital ad spending in 2017 will be 38.4% (77.37 billion USD) of total media ad spending. It will surpass TV ad spending which will be 35.8% of total.
  3. Google’s US advertising revenue is 31.00 billion USD in 2015, calculated from 67.39 billion global revenue of which 46% comes from the US. This is close to half of total digital ad spending (77.37 billion USD as noted above).
  4. Facebook’s 2015 advertising revenue was 17.08 billion USD. This is roughly a quarter of Google’s.
  5. As noted by Horace Dediu, economic growth in developing nations is not accelerating Google’s revenue growth. Despite rapid economic growth, developing nations are not becoming a larger part of Google’s revenue.

Assumptions

I will also assume the following;

  1. Google will not find a new revenue source that will be large enough to significantly add to its top line.
  2. Google’s revenue growth will continue to be dependent on and on par with growth in the US.

Logic

  1. Since the size of total media ad spending is constant as a percentage of GDP, this is the hard ceiling of advertising growth in the US.
  2. Digital ad spending is rapidly approaching this ceiling. With already close to 40% of total ad spending, there is less and less room left for digital to grow.
  3. Google has close to half of total digital ad spending. Of the remainder, it is likely that Facebook is taking half of this. Google has little space to grow by increasing its share within the total digital ad market. In fact, it is more likely that Facebook will eat into Google’s ad market share. Note that one estimate suggests that Google & Facebook own 85% of the US the digital ad market.
  4. Since Google’s ad revenue growth has largely been independent of developing countries, it is reasonable to assume that this will continue for the mid-term.

In simple terms, there is no longer room in the advertising industry for both Google and Facebook. Since Facebook has more momentum, it is likely that we will see Google being increasingly squeezed. Although the total digital ad spending will likely still see mid double digit growth, Facebook will take the majority of this growth and Google will probably drop to single digit growth before 2020.

What to expect in the future

We are already seeing signs of more disciplined spending at Google/Alphabet, most likely in anticipation of a slow-down in growth. Given the highly talented people at Google, it is no surprise that they understand that the end of double digit ad revenue growth is near.

However, disciplined spending can significantly alter what projects companies chase. Unlike the current Google which constantly throws spaghetti on the wall, a fiscally disciplined Google would probably be more cautious. Within the next few years, I expect that we will see a very different Google from what we are seeing now.

Update

One important thing to note is that “Peak Google” will be a result not of any strategic mistake made by the company, but rather a result of the saturation of the digital advertising market. This has the following implications;

  1. The whole digital advertising industry will suffer along with Google. In fact, smaller and less established players are more susceptible to adverse environments. This is already happening

  2. The saturation of the digital advertising industry also means the saturation of the ad-driven Internet. Startups without a monetisation model will find it harder to bolt-on an ad-driven one later.

  3. Being the most established brand in digital advertising, it is likely that Google will maintain a very strong position in the market for years to come. Like Apple, the issue will be the lack of rapid growth.

Smartphone Sales Down In Japan, But Android Hurting Most

MMRI released their report for mobile phone sales in Japan for he first half of 2016, and the results were not good. 

  1. Total handset sales decreased by 10.9% YoY. 
  2. Smartphones decreased 8.4% YoY. 
  3. However, looking at iPhone sales, this decreased only 3.1%, resulting in an increased market share of 40.7% (including feature phones).
  4. Importantly, Sony which is 2nd in market share saw a 28.5% drop in sales, while Sharp which is 3rd in smartphone share fell off a cliff with a 46.4% drop. 
  5. The sharp decline has been attributed to the government decision to restrict what they consider excessive discounting of devices. The government thinks that by restricting discounting (some smartphones are sold for free by carriers if the purchaser agrees to a 24 month contract), carriers will eventually reduce the prices of their data plans. However, data plan prices have yet to come down, and are actually increasing depending on your usage pattern. 

What this suggests is that when customers are more exposed to the real price of smartphones, it is the Android users who either decide to buy cheaper devices, or hold on to their devices longer. The iPhone users seem to be less sensitive to price increases. 

In a nutshell, the Android market has a high level of price elasticity whereas the iPhone market does not. 

I believe that the iPhone markets and the Android markets are actually different despite both being smartphones. Customers buy each for different needs, and they are not interchangeable. This is similar to how Mercedes and BMWs do not share the same market as cheap cars; the role of luxury cars is not just transportation.