Naofumi Kagami

Samsung Earnings 2Q14

Samsung’s earnings are in, and they were not very good (Bloomberg: note the original title in the URL slug. The title and tone of this article were revised from the pessimism in the original publication to a much more optimistic tone, which shows how worthless financial reporting has become).

Given these poor results, the blogosphere is awash with analysts telling us that they saw this coming. This is hardly something to brag about because it was so plainly obvious.

It’s also very easy to blame this on what Samsung does not possess; that is software and services. This is equally trivial to do. Giving a sound explanation of why Samsung is now in trouble, does not suggest that you have a clue about what is happening. Anybody can do that. It’s like saying the reason the New York Yankees are pretty weak right now is because their batters aren’t hitting the ball well.

To suggest that you really understand what is going on, you need to provide a coherent explanation of why Samsung was very strong a few years ago, and why it is now starting to fail. This explanation must consider that Samsung itself is mostly doing exactly the same things that there were before. Hence this explanation should not focus on factors that are internal to Samsung, but are external.

Hence the following explanations fail to suggest that the analyst who made them has a clue;

Marketing as a reason for Samsung’s success: Strong marketing is no doubt a reason for Samsung’s ascent to dominance. However Samsung has not relaxed its marketing focus. Unless we have a good explanation of why Samsung’s marketing is no longer effective (external factors), it doesn’t explain the situation.
Software & Services: Samsung is weak in software and services, but it has always been like this. In fact, Samsung knows this and has invested heavily into making itself better in software & services. The lack of strength in software & services does not explain Samsung’s problems because Samsung rose to dominance without them in the first place. Instead, we need an explanation of why software & services matter more than they used to or why Samsung’s efforts have become meaningless.
Cheap Chinese vendors: There have always been cheap Chinese vendors. Hence their presence itself does not explain anything. What has changed is that the Chinese vendors are now capable of creating products that rival Samsung’s in both quality and price at the lower-end of the market. What we need to understand is why the Chinese can now do this whereas they couldn’t do so before.

Clayton Christensen’s theories explain all of these. That’s why I lean on them.

In this context, what is interesting in Samsung’s report is actually the following;

Profit at Samsung’s chip unit, which supplies its own devices and also rivals such as Apple, nearly doubled to 2.1 trillion won on sales of 9.5 trillion won, according to the Bloomberg News analyst survey.

According the “the law of conservation of attractive profits”, as smartphone production becomes more modular allowing new entrants into the market, the attractive profits will shift to somewhere else in the value chain. Although not certain at this point, it is likely that component suppliers will be one of the stages where the attractive profits will accumulate. If so, Samsung’s chip unit may profit (although there is intense competition here as well).

It will be interesting to see how the component industry (Samsung’s chip business, MediaTek, etc.) does in the next few years.

Android L Screenshots

Uzair Ghani has put up screenshots comparing Android “L” preview to Android KitKat and iOS 8 beta.

A few notes;

Android L is definitely not yet finished. Even the calendar app uses the old Holo theme. Compared to where iOS 7 was when it was announced at WWDC, there still seems to be a lot to do. Maybe Google is planning to gradually update the Google Play apps; they may not yet be updated at the Android L release.
Android L does not seem to use the “frosted-glass” effect that iOS 7 has had (and iOS 8 also). On the iPhone 4, iOS does not use “frosted-glass” either. Maybe the GPUs are not yet powerful enough on most Android phones. This is a bit interesting because a lot of the UI effects on Android L look like they would depend on a good GPU.
The settings screen is no longer white letters on dark background. This is an important improvement in my opinion, as it sheds the final traces of Android’s geek-oriented history. It makes the setting panel much more approachable.
I’ve always been concerned about the general disregard for consistency in Android user interfaces, even in the Apps that Google itself designs. Unfortunately, we can already see this in the few screenshots. Specifically, in the L phone app, we see the action overflow menu on the upper left. It should be on the right. The action overflow has always been abused in Android, starting life as a hardware key. It has always been a UI control with low discoverability. It would be sad if it is still not getting the respect it deserves.

Android OEMs and The Law Of Conservation Of Attractive Profits

A couple of days ago, I wrote a comment on an article on the Tech.pinions website.

The article (by Jan Dawson) was discussing how Samsung came to dominate the Android OEM market, but how it is now struggling with competition on the low-end and is also having difficulty in differentiating itself.

The comment that I wrote;

It’s very interesting that you call Samsung “exceptional”. This is because, at least in my interpretation, the trajectory of Samsung’s accent and decline is precisely what Clayton Christensen’s theories would predict.

I’ve been thinking more about what I wrote, and I think that this is such an important issue that I should put it up as a blog post. Here, I will copy my original comment and add a bit more to it.

The rise and fall of Samsung as a smartphone OEM

Jan Dawson lists the keys to Samsung’s success as the following;

“On the marketing side, Samsung has vastly outspent all other Android smartphone manufacturers and become the default option for people in mature markets looking to buy a mid to high end smartphone.”
“Its vertical integration has allowed it to compete very efficiently and effectively with screen and other component technology.”

Jan Dawson also outlines Samsung’s current problems;

Overall smartphone growth is slowing, putting pressure on Samsung’s other device categories to provide stronger growth

Samsung’s dominant position in Android is being assailed at the low end and in the mid market by a variety of competitors, many of them from China

Google is reining in Android and looking to reassert its own position and services in the smartphone market, putting pressure on Samsung and others to tone down their customizations. New flavors of Android for wearables, the car and TVs will provide even less room for customization

People are at any rate apparently tiring of Samsung’s customizations of Android and starting to look more seriously at smartphones which provide a stock Android experience or at least something more like it

Samsung’s marketing spend is starting to experience the law of diminishing returns, where each dollar of spending no longer conveys the advantage it once did. It has effectively saturated the market and can no longer derive the advantages it once did from its far superior ad spend.

I completely agree with Jan Dawson’s assessment of why Samsung was successful, and Samsung’s current predicament.

The problem is, how did Samsung transition so quickly from “huge success” to “quite problematic”.

Jan Dawson doesn’t go into explaining the transition in depth. That is what my comment tries to do by applying Christensen’s theories.

How the environment changed from favorable to hostile for Samsung

Christensen’s theories are build on the premise that continued innovation (sustaining innovation) will ultimately open-up the possibility of low-end disruption, and that it can be very difficult for some companies in some markets to counter-attack the disruptor.

Hence to understand what happened to Samsung, we should analyze how technical advances changed the competitive landscape, and allowed low-end disruptors to enter the market. That is what I tried to do in the comment;

When the product is not good enough, the attractive profits flow to the integrator. This was the situation at the beginning. Samsung’s ability to integrate the hardware stack and to also put a UI (that was attractive in the sense that it more closely imitated iPhone) on top was the reason they had the best Android product.

As technology improved, customized integration became less necessary. This caused modularity in the hardware stack. SoC vendors like MediaTek are prime examples of the hardware becoming more modular. Also, as Android got its act together and became less ugly, Samsung’s ability to put their UI on top became less important and even downright annoying. Hence the software vendor (Google) increased its power. In fact, integration within the software stack increased. In total, the Android value chain became modularized and Samsung’s strength as an integrator waned.

Samsung tried to buck this trend by creating products that we much better than those assembled from modular parts (which is the same as Apple’s strategy). Hence they designed their own CPU and added features to their software. Unfortunately, both were unsuccessful. Neither created value that appealed to their customers.

The first item explains why Samsung rose to dominance. Samsung is very vertically integrated in hardware. It has top-level semiconductor technology, display technology, wireless technology, etc. Importantly, because Apple had to rely on Samsung for key components, Samsung learned of Apple’s iPhone plans many months before their competitors. None of their competitors had a similar advantage. This allowed Samsung to rapidly produce the best Android smartphone hardware

One thing I want to add that is rarely mentioned is that Samsung’s TouchWiz UI was well received during these early years. Before Android 4.0, the Android UI was pretty bad (Android 4.0 was a huge overhaul of the UI). Samsung’s TouchWiz imitated the iPhone UI much more closely than stock Android, and this was no doubt one of the reasons why Samsung was so successful. Hence in this context, Samsung’s ability to create a good UI skin was also a key factor in its success. As far as I can tell, yearning for a stock Andorid experience is post Android 4.0. Until then, stock Android was not “good enough”.

Going on to the second item, we start to see how technical progress changed the environment. Christensen has described how markets eventually favor modularity over integration after the products become “good enough”. As component technologies improve, it becomes less important to fine-tune the components to obtain the necessary performance. Customers are now satisfied by products that are simply assembled from “off-the-shelf” components (from SoC vendors like MediaTek for example). In many cases, even the assembly is outsourced to China. Because fine-tuning is less important, this allows low-end entrants without extensive hardware expertise to enter the market. Because the low-end entrants do not need much hardware expertise, their business models are often very different from Samsung’s. For example, Lenovo is a low-cost assembler of hardware, an operation that doesn’t require much R&D spending. Cherry Mobile, a fast growing smartphone brand in the Philippines is actually a carrier, not a hardware maker. Xiaomi, a fast growing fabless smartphone maker in China that sells mid-range phones with razor-thin margins, earns profit not through hardware sales but through service sales and reduced costs on marketing.

It is also important to note that many of the new entrants mainly have strengths in their local markets. Cherry Mobile (Philippines), Xiaomi (China), Lenovo (China), Micromax (India), Wiko (France), BQ (Spain) are some of the examples. Being local companies, they have some marketing advantages over Samsung and can better tailor their products to local preferences.

To reiterate, technical progress has allowed low-end entrants with a very different low-cost business model, to enter the smartphone market. Samsung’s integration and expertise in hardware is no longer a competitive advantage. In fact, it could even become a liability because it could make it difficult for Samsung to directly address these low-end entrants.

Technical progress has changed the initially favorable environment for Samsung into a hostile one.

Google’s increased power over OEMs

This is explained by “the law of conservation of attractive profits”, which I have described in a separate post.

It also explains why Google now has the power to reign-in Android. The negotiation power of the integrators has declined and has shifted to the OS and service vendor. Google did not have this power until the hardware stack became modular.

The law of conservation of attractive profits allows Google to have stronger influence over Samsung. However, as I described in the previous post, I tend to think that this is temporary and that the attractive profits will transition back to the OEMs who have alternative business models.

Looking into the future

One of the more interesting features of the recently announced Android “L” is the new ART runtime. This looks like it will significantly improve Android performance on low-end hardware. This will accelerate the shift in power away from Samsung and towards the low-end entrants. Combine this with the Android One reference platform, and we will likely see almost all value in low-end Android hardware sucked out. Samsung will have a very hard time competing in this market, and since their competitors are now asymmetric, it’s very hard to see how they can launch a strong counter attack. However, since most of the entrants are local, the spread of low-end entrants might be a gradual process.

On the high end, especially in markets where the cost of the smartphone is subsidized by the carrier, Samsung will continue to be the dominant Android OEM. Other Android OEMs simply cannot match Samsung in marketing and product features. Whether that is enough given the strength of iPhone is yet to be seen.

Importantly, Google’s strategy is providing absolutely zero incentive and zero profits for OEMs to innovate in hardware. With this market structure, it’s difficult to see how the low-end entrants could evolve to the point where they can overtake Samsung in the high-end, even Xiaomi or Lenovo. I don’t see this ever happening in the foreseeable future.

Actually, this could make Apple more dominant in the high-end, and Google probably won’t even care.

Mobile Money

Leo Mirani recently reported on how mobile money transactions are becoming popular in emerging markets. It’s an old story, but it is increasingly relevant as Google falters with Google Wallet and Apple is rumored to enter payments using Touch ID fingerprint scanning.

I’ll just list a few of my observations and thoughts.

The technology for mobile money in emerging countries is very low tech. In fact, it is based on simple text messages.
The US in particular is very slow to adopt mobile money, instead relying on credit cards (and sometimes paper checks).
Although not as spectacular as countries like Kenya and Tanzania, Japan also has quite popular payment systems that are reliant on mobile.
What we are witnessing is a general phenomenon. Countries acquire habits as they grow their economy and these habits persist as they mature. The technical and cultural environment during the growth period strongly affects these habits, and they can be difficult to change later even after environments change.

Africa is growing its economy in an age where almost everybody has a mobile phone. That’s why mobile payments are popular. On the other hand, the US credit card system emerged when magnetic cards were still high-tech (remember the days when we passed credit cards through a carbon copy roller?). Although almost everybody in the US now has a mobile phone, it’s difficult for the US to change its habits despite the benefits that it would bring.

Japan, happily enough, is one of few countries where you could carry large amounts of cash without worrying about being mugged. Because carrying cash was not a concern, credit card adoption was slow and was incomplete. Hence it was possible for mobile payments to thrive in certain niches.

It’s really simple to understand how each country is the way it is. However, it’s very difficult to change.

What this means is that each country has very different requirements for mobile payments. A one-size-fits-all approach is destined to fail. Working with local partners and customizing for the local situation is essential.

Returning to the question of how Google and Apple should approach mobile payments, it’s very clear that trying to be the payment system will only work for a certain market segment. It is more important to empower the system that each country already has through APIs or maybe something like HealthKit. Helping others to succeed on your platform and benefiting indirectly should be the way to go.

Gmail with Material Design

I was just thinking about Material Design within Google’s own apps, and it occurred to me that at least in my case, the Gmail app would never look like the screenshots on the web.

For example, the screenshot below was taken from CNET and it surely looks nice (Material Design on right, current design on left).

NewImage

However, my Gmail account on my Android phone actually looks like this (yes, I only use my Gmail account to sign up for newsletters and mailing lists. I never use it for communicating with someone I know, but you get the idea);

Screenshot 2014 07 02 20 13 05

When the current Gmail cannot find a photo, it uses the first letter of the sender’s name instead. This is totally ugly, and what’s more, completely useless.

I wonder how Material Design solves this…

Or maybe it’s a setting deep in the menus…

I think we have to keep in mind that when we communicate with close friends, we tend not to use email anymore; we use messaging services, Twitter or Facebook. For these accounts, we often have photos. Emails are for work, mailing lists, notifications, etc. We normally don’t have photos for these. At least, I don’t.

To come to a good design, designers have to understand how the product will be used, in what context and with whom. At least, I think that is how good designers are supposed to approach problems.

Statistical Validation of Disruption Theory

There has been a bit of discussion on the web about the validity of Clayton Christensen’s Disruption theory. I hope this article by Thomas Thurston will put it all to rest; “Christensen Vs. Lepore: A Matter Of Fact”.

Most people don’t know this, but it turns out Disruption Theory is the foundation of the most accurate, thoroughly vetted, quantitative prediction models of new business survival or failure in the world today.

…

I did my best to reduce his theory to falsifiable yes/no logic using published research. Even so, in the first round these relatively crude rules based on Disruption Theory blindly predicted if new businesses would survive or fail with 94 percent accuracy and over 99 percent statistical confidence. Holy crap.

So statistically, disruption theory is pretty good a predictions.

How often does it get it wrong?

A lot of people point to examples of when Disruption Theory, or Christensen, was wrong. It was wrong about the iPhone. Tesla. Ralph Lauren. In fact, it’s been wrong over 7,500 times by my count (remember it has a 33 percent error rate when predicting winners). Keep in mind, however, it’s 66 percent right while everything else is stuck at 25 percent. Improvement, not perfection, is the standard.

Well, it doesn’t get it right 100% of the time. Does that mean the theory isn’t valid? Not at all.

Many bloggers dismissed or attempted to modify Disruption Theory because it got Apple wrong. These people don’t understand statistics or don’t understand how to use statistics to validate or invalidate an argument. I could list them here, but I won’t.

The method used here is very standard. For example, every new drug that comes to the market is tested for effectiveness using the same statistical methods. Drugs sometimes work, but often they don’t. All they have to do is work a bit better, a bit more often. If we rejected every drug that ever failed to work for certain patients, there would be no drugs.

I especially like Thomas’ closing sentence. It’s the standard way of statistically testing which theory is correct, which Thomas uses to test Lepore’s theory against Clay’s. Anybody who doesn’t fully get it shouldn’t be talking about how theories are validated.

Lepore could be right about Disruption Theory, but the odds are literally over 500,000 times greater that, as a matter of fact, she’s just plain wrong.

P.S.

Although I am appalled that a large number of people who discussed Disruption Theory on the web didn’t seem to understand the basic principles of statistical testing, I understand where they came from.

Statistics simply isn’t taught enough at schools. Statistics, in my view, is one of the most important branches of mathematics and is relevant even for people who won’t touch maths ever in their professional careers. We have to understand statistics so that we, in democratic nations, can correctly assess the accuracy of claims made by politicians, and so much more.

This is a real shame.

Google Self-Driving Cars and The Car Industry

Reuters article; “Google, Detroit diverge on road map for self-driving cars”

Not too much of a surprise for anybody who has worked in a regulated industry.

In 2012, a small team of Google Inc engineers and business staffers met with several of the world’s largest car makers, to discuss partnerships to build self-driving cars.

In one meeting, both sides were enthusiastic about the futuristic technology, yet it soon became clear that they would not be working together. The Internet search company and the automaker disagreed on almost every point, from car capabilities and time needed to get it to market to extent of collaboration.

I think that this part sums up the situation well;

“There was a certain amount of arrogance on the Google side, in the sense of ‘We know what we’re doing, you just help us,’” said a second person, representing a major car maker, who was involved in discussions with Google.

“We’d say, ‘Well you don’t really know that much. And we’re not going to put our name on a project like that because if something goes wrong, we have a lot more to lose.’”

I seriously doubt that Google has the patience and the willingness to work with others that is required to pull this off.

Japan’s Largest University Switching to Microsoft Office 365 from Google Apps (Docs)

The largest university in Japan, Nihon University will provide “Office 365 Education” for all of its 100,000 students according to the Microsoft Japan website.

I’m still in the process of researching the details, but some things that have been mentioned that I find very interesting;

Nihon University had been using “Google Apps Education Edition” since April, 2007.
Reason 1: Faculty staff used Google Apps but student uptake was not good. Students preferred to use their own free mail accounts.
Reason 2: Unfamiliarity with Google Apps was a reason for slow uptake. By providing the software that everybody is familiar with (Office), Nihon Univ. hopes that students will also use scheduling and address book features.
Reason 3: Students were pirating MS-Office install disks. Office 365 will make that a non-issue. Nihon Univ. chose the A3 plan with Office 365 ProPlus, which means that faculty and students can install Office on 5 PCs per user.
Annual price per faculty is 410 JPY, per student 230 JPY. This ends up being cheaper then when they were using “Google Apps Education Edition” because even when they were using Google, they still needed to buy significant installations of MS Office.

Although we still need more examples to see whether this is a trend or not, I sense strong beginnings.

Cloud is getting cheap

The most powerful allure of Google Apps is the price. For general consumers and for education, the price is free. This was possible because Google had a robust advertising model. By injecting ads in the web user interfaces, Google could justify the cost of providing the service for free.

Historically, Google was uniquely positioned to provide an office suite for free. Other companies could not do this profitably.

However, as technology improved and the hardware required for cloud services dropped in price, it became feasible for companies without a robust advertising model to provide free or very cheap cloud services. This can be witnessed in the recent announcement at Apple’s WWDC 2014. Apple announced that they would be providing CloudKit effectively for free.

We are now at the point that we don’t even need advertising anymore. It has become feasible to provide free or very cheap cloud office suites, even without advertising. Hence anybody can do it. Google no longer has a unique advantage in providing services for free.

Google Apps never became “good enough”

With the cost advantage of Google Apps eroding, the argument for choosing either Google Apps or Office 365 now rests on the benefits that each platform provides. This is something that Google Apps was never designed for.

Since its inception, Google Apps was designed as a simplified version of MS Office that justified its existence by being much cheaper. Although it had some unique collaboration features, it never evolved to become better than MS Office. It was always obvious that if it lost its cost advantage, it would lose out against MS Office.

Looking at the reasons why Office 365 was chosen over Google Apps, it’s very apparent that Office was still an application that both faculty and students needed to use from time to time. Google Apps had never become “good enough” on its own.

Microsoft is changing

Since the costs of providing cloud services had decreased, the only roadblock for Microsoft going aggressive with Office 365 was the possibility of cannibalization. Office 365 could potentially cannibalize sales of their standalone office suite.

A few thing have happened that might have changed their minds.

Adobe has been very successful with their Adobe Creative Cloud.
Microsoft’s new CEO is a cloud guy.

What are the trends?

These are the trends that I think we are beginning to see.

Google’s strength in advertising will no longer be sufficient to maintain an advantage (based on cost) in the cloud. Cloud costs have gone so low that advertising is no longer necessary for a free/low-cost service. Subscription services are proving to be a good business model.
With the price issue becoming less of a concern, competition in the cloud will focus more on features and usability. In established markets, one feature that will continue to be extremely important is compatibility with de-facto standards (both in file-formats and user interface).
As the focus shifts to features and usability, native applications will maintain their advantage against web apps.

Chinese Android Larger Than Google’s Android

Chinese Android, that is the Android that is based on Android Open-Source (AOSP) and is independent of Google’s services and restrictions, seems to be already significantly larger than Google’s Android (the one that relies on Google services, and comes with Google’s restrictions).

At least that seems to be the case if we look at app download statistics.

According to Analysys International, App downloads in China were 23.4 billion for 2014Q1 alone. In comparison, Google Play downloads were probably in the 15 billion range for 2014Q2 (Estimated from 1 & 2). That’s quite a big difference.

Revenue-wise, I don’t have any data. However, given that Google Play revenue is dependent on mature developed nations (Japan, US, South Korea), it is very possible that revenue is growing much faster in China. Furthermore, iOS App Store revenue has been pretty high in China, suggesting that the Chinese have a rather high propensity to spend money on Apps. I would not be surprised if Chinese Android app revenue is similar to or has already surpassed Google Play app revenue.

Of course this discussion hinges on the “23.4 billion for 2014Q1 alone” report being true. Hopefully, we will get verification soon.

Obviously, the implications of this are huge for the Android ecosystem.

Material Design On The Web: A Well Funded Research Project

The new Material Design announced at Google I/O is supposedly an ambitious cross-platform design language that spans not only Android devices, but also the web.

Anything that is this ambitious should be treated with a healthy grain of salt. Let’s verify what Google means by cross-platform.

Android version compatibility

Given the severe OS version fragmentation of the Android platform, it is important to understand the entirety of version compatibility for Material Design.

Unfortunately, I could not find resources on the web that gave me a good comprehensive idea. I think Material Design will only be supported on Android L. There appear to be some support libraries that will make it easier to support older versions and a Material Design version with the same code, but it looks like that will be it.

Browser compatibility

Google does provide details of browser compatibility for Material Design on the web (the Polymer project). Unfortunately, it doesn’t look too good. According to this discussion on the web, Google is not taking browser compatibility very seriously.

In terms of compatibility, it is simply not up to the standards that most web developers would regard as sufficient.

Summary

The lack of Android version compatibility is understandable and will not present an issue. It will be adopted over time.

The issue is with Browser compatibility. Although impressive, there are many similar projects to bring a nice UI to web apps, and none are as restrictive in compatibility as Material Design. Web developers can simply chose one of these. There is little reason to believe that web developers will slowly adopt Material Design.

It’s still very much a research project.