Amazon – Naofumi Kagami

Smart Speakers and Clocks

Traditionally, most rooms in a house have not had information devices in them. Some rooms may have a TV and some may have a radio. However other than that, most rooms do not have any installed equipment that enables the residents to quickly get information.

There is one exception however, and that is clocks. Many rooms have clocks either hanging on the wall or sitting on a mantelpiece. Clocks provide information that is vital for modern life (“what time is it?”) with a mere glance, and they are so useful that people put them in their bedrooms, living rooms, kitchens, home offices, and sometimes even in their toilets.

I wonder if the best way to think about Smart Speakers and the utility that they may bring to the home, is to compare them to these clocks. Can Smart Speakers provide the same benefit that we get from clocks around the house? Is the information that Smart Speakers may provide as important as telling the time? Why do we still need wall clocks when we can easily wear wrist watches instead? Does the UI need to be distilled to a simple glance instead of a voice UI?

If you adopt this viewpoint, then I think you can naturally arrive at the following conclusions.

People generally don’t need too much information in the house. The time is a rare case of something that we often need to know in multiple rooms.
If the information is important and frequently needed, then people will want to put devices in multiple rooms instead of relying on a watch or a smartphone that they carry around with them.
People will prefer to get the information through a simple glance. A voice UI may be too inconvenient. For example, instead of a smart speaker that can tell you the weather when asked, a wall clock with a display showing this will probably end up being more useful.

Amazon’s Echo Strategy

I have read over the Internet that some pundits are estimating that Amazon sold tens of millions of Echo devices – much more than their closest rival, Google’s Home.

Impressive as this may sound, it is not the first time that Amazon has used a strategy of dumping a cheap/subsidised device onto the market with the goal of winning market share early on in the game. The original Kindle Fire tablet that was introduced in 2011 shared the same strategy and was also a hit during the holiday shopping season, topping Amazon’s best seller charts as well. The Kindle Fire never became dominant though.

The lesson that I would learn from this is that flooding the market early on with cheap devices will not win you a strong position for the future. The tech market constantly evolves and products reliably get better and better each year, almost like clockwork. Even though smart speakers may look completely cloud-dependent with very few requirements for local hardware, I can reliably predict that in the next few years, this will no longer be the case. If the market for smart speakers persists (which is by no means a given), they will for example at least evolve to incorporate some local AI features to allow them to become smarter while maintaining a certain level of privacy. Cheap devices that are a few year’s old and do not have these improvements will not provide any kind of significant moat, and customers will be eager to switch to the new ones.

Predicting Technology Adoption

Today we see many new ideas and product categories being introduced into the market. We see smart speakers, AI assistants, self-driving cars, wearables, VR headsets, smart homes and many more. Not all of these will be successful and become a part of our future lives. Surely some eventually will, but only after many years of mediocre sales mainly as an enthusiast’s toy. Others will become instant hits, and will be used by the majority of people within a decade or even shorter. Despite all of these sharing a common early enthusiasm, the eventual fates will be very different, and predicting which trajectory each product category will take is not an easy task. However there are some frameworks and theories which we can apply, so here, as an excercise to test my own understanding, I will try to do just that.

Smart Speakers

This is a very interesting product category to test and apply the Chasm theory of Geoffrey Moore and the Disruption theory of Clayton Christensen.

The Chasm theory applies to new category products that require customers to learn and change their behaviour. Essentially, the Chasm theory states that there is a very wide divide between the early adopters and the early majority, due to the lack of connection between these two different types of customer. The early adopters do not act as reliable references for the early majority, and hence technology diffusion cannot rely on word-of-mouth. Thus a deliberate and focused strategy is required to bridge this gap. On the other hand, Disruption occurs when the core appeal of a preexisting product that caters to advanced users is made available to a wider audience through technology or business model advancements. As a result the market share of the preexisting product is significantly reduced (hence the word Disruption), often not necessarily by stealing the preexisting product’s customers, but instead dramatically increasing the size of the pie. Often in Disruption, the benefits of the preexisting product will already have been defined and will be very apparent to the new users, and so there will not be a Chasm to cross. This can make adoption very rapid.

In the smart speaker market, and with the introduction of Apple’s HomePod, we will witness both of these theories simultaneously at work.

For the Amazon Echo and Google Home products, we will see the effect of the Chasm. Both of these products stress the AI assistant features which are new and without precedent in the market. Therefore the companies will struggle to define the must-have value proposition, and to educate consumers as to why they should own one. Geoffrey Moore advocates a whole product approach to overcome this, something that neither company is particularly known for. I therefore predict that both these products will seriously struggle to be widely adopted.

On the other hand, we see Apple’s HomePod taking a more Low-end Disruption approach. Instead of touting AI assistant capabilities, the HomePod takes aim at high-end audio and making that accessible to people who would appreciate high-quality sound, but who are unwilling to invest in the learning and the equipment. Therefore, this approach will not experience a Chasm. Potential customers already know what they are getting and do not need to be educated on why they would enjoy good sound. All they need to decide is when they are going to save up, and therefore we are likely to see a much faster adoption rate with this marketing focus.

As a result, my prediction is that the HomePod will be a hit akin to the AirPods, easily surpassing Amazon Echo sales in revenue and even units. As a result we will see both Amazon and Google changing direction and copying Apple’s approach by making speakers with at least acceptable sound. The smart speaker market will rapidly expand as a result, but only because the marketing approach is not about the AI assistant capabilities but the high-quality music.

Self-driving cars

Self-driving technology is also likely to follow a Chasm trajectory. Without a preexisting product with a similar value proposition, people will have to be educated on the benefits, especially if regulations require that the driver cannot fully delegate the role of driving to the AI, and must be available to intervene at any time. We will see a lot of early adopters, but for a long while, that is likely to be all. Even if full autonomy becomes a reality soon and regulations allow the computers to work without human supervision, I am not convinced that many people will immediately recognise the benefits since there is no preexisting market.

Wearables

This is product category that we know is in the Chasm. The benefits of the Apple Watch for example have been hard for even Apple to define, and consumers have yet to be educated of the benefits. Apple is certainly the most experienced company when it comes to crossing the Chasm (it was the company that succeeded in popularising the mouse and the GUI), but still it will take time.

The problem for wearables is that there is no preexisting product to disrupt. For example, except for a very limited number of people, tracking your heartbeat is hardly necessary regardless of how expensive or complicated the devices may be. Notifications on your wrist is also something that most people cannot immediately grasp the benefit of. On the other hand, it is also obvious that with only a minuscule screen size, a watch cannot replace the current day smartphone, no matter how powerful it may be or even if it is connected to an LTE network. Without a preexisting product, it’s very hard to understand what a wearable is good for. It is difficult to understand what benefit it can provide to a customer.

Remember that Apple initially positioned the Apple Watch as a fashion item and a high precision timepiece. The significance of this strategy was that they adopted the value proposition of a preexisting product, which enabled them to connect to customers who were not typical early adopters, thereby speeding up adoption and leaving Android Wear in the dust. Also note that while the regular smartphone manufacturers like LG have recently decided to wind down their smartwatch efforts, traditional watch vendors have started to do the opposite. Regardless of whatever tech is included in the product, at this early point in time where most consumers have no idea what tech can do for them, the value proposition of SmartWatches remain mostly the same as traditional ones and therefore the traditional vendors are the ones that have a story to sell.

Smart homes

Again, this is a product category with no predecessor. Even if you were rich enough to employ a maid or a butler, I suspect switching on the lights would not be typically something that you would do via voice commands. What has been much more critical and aligned better with what people typically hired maids for, was cleaning up and vacuuming the floors, something that Roombas do quite well without any need for a voice controlled “smart” UI. This is why we have seen pretty rapid adoption of these vacuum robots whereas Smart homes are still very much an enthusiast’s toy.

Smart homes will predictably hit a Chasm. The benefits are not clear and therefore only early adopters will buy these things. This will continue to be the case for at least several years. Early majority customers will not understand the benefits without a whole product strategy around a key use case, whatever that might be. Adoption will take a long time, if ever. On the other hand, we will continue to see innovations that use computer, sensor and software technologies to make household chores easier and quicker to do. It’s just that switching on lights is not one of those chores that is a recognised burden on our day-to-day lives.

Summary

The key framework that I’ve used above is to first identify if there is a precursor to a new product category and to see if that has already made consumers aware of the features and benefits. If so, and if the new product has the potential to dramatically increase the market, then I predict that we will see rapid adoption (whether its will be Disruptive or Sustaining depends on whether the incentives to pursue the new product aligns with the business models of the incumbents). Otherwise, we can expect the new product category to follow the technology adoption life cycle and to hit the Chasm. Since the Chasm can be overcome with a whole product strategy, whether or not the market players have experience in this is key to adoption speed.

With this framework, we can predict the rapid adoption of smart speakers by virtue of Apple’s high-quality music strategy, the relatively good adoption of smart watches due to Apple’s experience in whole product strategies and marketing, and the slow adoption of self-driving cars & smart homes due to the current lack of either.

It is also important to note that the key players during the early adopter phase are not necessarily the ones that will make it through to the early majority phase. The Commodores and Amigas of the early PC market did not thrive when PCs became mainstream. Likewise, it is a fallacy to assume that the companies that are currently “winning” in smart speakers and self-driving cars or even Artificial Intelligence in general will enjoy their early lead as the market goes mainstream. More likely than not, other companies that are better positioned for the early majority market and the specific use cases will take it from them.

Out of the markets that I have described here, the smart speaker market is the one that is most likely to see significant action in a year or two due to Apple’s fresh approach. I expect to revisit this post and review my predictions mid or at the very latest late-2018. Other ones will probably still be stuck in the early adopter market and it will be harder to say whether my predictions were right or not.

Why Do Companies Outsource?

Last week, news broke out of Apple’s “McQueen” project; a plan to move Apple’s iCloud data away from Amazon’s AWS, Microsoft’s Azure, etc. and into its own data centres.

This isn’t really a surprise or an insidious plan by Apple to damage its competitors. It’s obvious if you think about what the benefit of outsourcing is for any company.

Companies outsource if;
1. The technology is not a core strength of the company.
2. The technology will not be a key differentiator going forward.
3. There are cost benefits (mostly due to scale) of outsourcing.

In the case of Apple, cloud infrastructure was not a core strength of the company so it made sense to outsource at the onset. However, it became clear that the cloud would be a key differentiator going forward. Additionally, the scale of Apple’s cloud operations became huge, and hence the cost benefits of outsourcing became negative.

The only thing that was surprising to me was the fact that Apple was outsourcing at all. I would have thought that Apple had had everything in house years ago.

Amazon and the Capitalist’s Dilemma

Jan Dawson wrote a great summary of Amazon’s business and gave some ideas on why Amazon might be buying Twitch.

I am very unfamiliar with Twitch, but Jan’s summary of Amazon fits some thoughts that I have regarding Clayton Christensen’s “Capitalist’s Dilemma”. I would like to touch on that here.

I previously wrote a very brief summary of the Capitalist’s Dilemma on this blog.

Not all innovations are equal. In fact, the majority of the innovations that are happening today are detrimental to economic growth. The authors dissect “Innovation” into three separate categories and argue that the one that create jobs (market-creating innovations) is currently being de-emphasized, while the one that eliminates jobs (efficiency innovations) is being highlighted.

Jan Dawson summarizes Amazon’s business as follows;

Essentially all of Amazon’s business rests on the transfer of spending from legacy categories to categories it competes in, whether that’s e-commerce replacing bricks-and-mortar retail or digital content replacing physical content (or even cloud computing replacing premise-based computing). As such, Amazon’s addressable market is directly tied to three factors:

the size of the legacy markets it’s seeking to disrupt

the degree to which those markets are shifting into categories Amazon competes in

the market share Amazon is able to capture.

From here, it is plainly obvious that Amazon’s business (with the exception of AWS perhaps) perfectly fits Christensen’s description of an “efficiency innovation”. That is to say, Amazon’s business eliminates jobs and is detrimental to economic growth.

Unlike Christensen’s “Innovator’s Dilemma”, the Capitalist’s Dilemma will not directly impact a company’s business. In fact, it will usually greatly benefit short term profits. Instead, it will slowly deteriorate the health of the company, leading to long term effects. Even more significantly, it will negatively affect the national economy as a whole.

This suggests that Amazon itself will benefit from its current focus on replacing brick-and-mortar retail. This may even be true in the long term. However, Christensen’s observations predict that the US economy as a whole will suffer.

P.S.

Using the Capitalist’s Dilemma as a lens, it is possible to argue that services like Uber are also mainly “efficiency innovations”. It’s definitely something to consider before singing praises of tech innovation.

More on Attractive Profits in the Cloud

I’ve been touching on the subject of commoditization of the cloud a couple of times on this blog(1, 2.

Today, I’ll like to look at the current players and how a possible commoditization of the cloud will affect their businesses. That is to say, do the current players own an adjacent layer in the value chain that can reap the attractive profits.

Barb Darrow at GigaOM, citing Rick Sherlund from Nomura Securities, gave some estimates on which companies are are making money from their cloud business (below). It is clear that Amazon AWS is losing it’s position as the dominant cloud vendor. The other companies, Salesforce, Microsoft, IBM and Google are significantly narrowing the gap. Microsoft and IBM in particular have very high growth rates relative to AWS.

NewImage

Given that cloud services are in many ways similar to rental servers, and that it is probably difficult to maintain differentiation, it is likely that prices will drop and cloud services will commoditize. When commoditization happens, Christensen’s “law of conservation of attractive profits” predicts that the attractive profits will shift to an adjacent layer in the value chain. Here I would like to see if the current players in this market will be able to capture the profits as they shift.

Microsoft, IBM: Both of these companies are very strong in enterprise IT. As the cloud commoditize and enterprises move their data centers into the cloud, Microsoft and IBM could easily provide value through consultation and customized services. They are both well positioned to take advantage of cloud commoditization.
Salesforce: Similar to Microsoft and IBM, Salesforce will benefit through consultation and customization for the cloud services.
Amazon: I can’t see any of Amazon’s strengths in the layers adjacent to the cloud. I don’t see them benefiting at all from commoditization of the cloud.
Google: Google makes almost all of it’s money from advertising, and it doesn’t make much money (at least not directly) from other activities. Hence it is difficult to say whether they are capturing profit or not in any of their activities other than search. Likewise, it is difficult to discuss how the commoditization of the cloud will affect them.

Amazon’s Hardware Strategy Mystery

Ben Bajarin recently tweeted a chart showing how tablet market shares have changed since 2011.

BenBajarin 2014 7月 25

In Ben’s tweet, he noted how Amazon’s market share has now become very small, and that is certainly true. Although Amazon’s share tends to rise sharply in the holiday season and we can still expect another bump in Q4 of this year, 14Q2 is seeing very very small sales for Amazon.

At this point in time, I think that it is worthwhile to review how Amazon started selling the Kindle Fire, and how its strategy has unfolded.

The Kindle Fire launched in 11Q4 at a very low price of $199 (compared to $499 for the iPad) which was only possible because Amazon subsidized the price.
The Kindle Fire was the catalyst that showed how Android tablets could take market share away from Apple. Up till then, companies like Dell and HP had tried to compete with Apple without any noticeable consequences. The extremely low cost of the Kindle Fire finally allowed Android tablets to start expanding market share, and it became obvious that Android tablets had to be priced in this domain if they were to ever sell well.
The Kindle Fire was a low-end device and omitted many features that were present in the competition.
The success of the Kindle Fire demonstrated for the fast time that Android tablets could compete against the iPad under the condition that the price was kept under 200 USDb (many Android tablets had previously tried to enter the market, but all had failed).
In particular, Google’s Nexus 7, which was released in July 2012, was obviously designed to match the successful formula that the Kindle Fire had pioneered; a 7-inch screen and a price below 200 USD. The Nexus 7 shows up in Asus sales in Ben’s chart, which was significantly elevated since 12Q3.
Amazon released the second generation of the Kindle Fire in 12Q4 with a price reduction to $159. They also released two higher-spec versions, the Kindle Fire HD priced at $199 and $299. With these updates, Amazon saw market share similar to what they saw with the first version.
Amazon released the third generation in 4Q13 for 139 USD, again together with two higher-spec models ($229 and $379). Amazon again saw good sales in the holiday season, but sales dramatically dropped in the next year according to Ben’s chart.
Importantly, the Nexus line of tablets (produced by Asus) also lost steam after the initial introduction.
The only branded tablet that saw an increase in market share during this period was Samsung. Although I cannot find pubic data that explains Samsung’s rise, it is likely that Samsung’s sales were the result of bundling with smartphones and possibly TVs; they weren’t selling by themselves and people got them even when they didn’t really want them.

We can clearly see that despite initial expectations, the Kindle Fire tablets have not really grown to be a strong contender in the tablet market. This is also true of the Nexus 7 lineup. The question is, could Amazon have done better?

We know that a large part of the “others” in Ben’s chart are non-branded tablets, which are generally very cheap, low quality and come out of the Chinese technology ecosystem. These have been selling mostly for watching video. Samsung’s sales also come from bundling which means that the cost to the consumer is very low or maybe even free. What this means is that although the Kindle Fire was just about the cheapest usable tablet when it debuted, that is no longer the case. Kindle Fire languished because it was no longer the cheapest tablet that was barely usable.

This suggests that Amazon could have done better with the Kindle Fire if they had continued to pursue their low-end strategy. Instead of moving up-market with the Kindle Fire HD and HDX, they could have instead gone lower into possibly sub-100 USD price points. They could have even provided the Kindle Fire for free with an Amazon Prime membership subscription. In fact, this is exactly the strategy I had expected Amazon would pursue. I actually wrote a blog post back in October 2011 (in Japanese).

In my old blog post, I expected Amazon to focus on creating a tablet that was so limited in features that the only thing you could do with it was to consume content. I expected that Amazon would continue to omit the camera, mike, gyroscope, and cellular connectivity, because these were not essential for reading books or watching videos. They could even have continued to use Android 2.3 as the base for their OS.

This is not the strategy they pursued. Instead of staying at the bottom, they moved up to the mid-market segment. Even though Amazon’s prices decreased, competitor prices dropped even faster. The Kindle Fire is no longer a product that stands out in price, and as such, it has lost its unique appeal. No wonder that sales are not expanding.

It’s a mystery to me why Amazon is pursuing the mid-market in hardware. If you look at the recently announced Fire phone, they are even trying to sell a high-end phone with a significant profit. My opinion is that aiming for the mid- to high-end does not make any strategic sense for Amazon. I am totally bewildered and I’m suspecting that Jeff Bezos is starting to get confused.

Design And The Lack Of Intent

I read a great post today by John Moran about Intent within Apple’s designs.

Overarching intent is easy. The hard part is driving that conscious decision-making throughout every little choice in the creative process. Good designers have a clear sense of the overall purpose of their creation; great designers can say, “This is why we made that decision” about a thousand details.

When Jony Ive, Apple’s newly titled SVP of Design, criticizes a material selection or feature decision, “he’s known to use ‘arbitrary’ as a term of abuse.”

John goes to outline the “Three Design Evasions”; what most companies do instead of employing Intent.

Preserving the past.
Copy first without making the Intent your own.
Delegating design decisions to your customers.

The question is, what is holding good design back? Do we lack good designers or are corporations ruining them?

I don’t know the answer. I’m quite sure that a lot of designers are aware of Intent and consciously try hard. The problem is, I tend to find most of the celebrated designers lacking it the most. Instead, I often find good Intent and good design in kitchen utensils and other everyday tools that are not “appearance” driven.

For example, architecture. Widely acclaimed architecture is more often praised based on its appearance. Of course critics will note Intent and usability benefits. However, as an actual user of these buildings, I have never found myself appreciating the designers’ decisions. In fact, we more often tend to loath the strangely designed rooftops that invariable leak rain, and the unconventional hallway orientations that make you feel lost.

On the other hand, I marvel at the curves of my kitchen cutting knife or scissors which are truly designed to fit your hand. I wonder at the small details on the metal knife embedded in the box of my food wrap which allow me to efficiently cut the film with the minimum of strength and without it tearing in unwanted places.

Even Jony Ive’s designs used to have annoying flaws. I hated the trackpad buttons on the Powerbook G3s which looked nice, but had a very badly designed clicking mechanism. The flimsy hinge on the Titanium Powerbook easily fractured and came off. And it’s really hard to justify the design of the “toilet-seat” iBooks. There was very little Intent in those designs.

In my opinion, it was only after the aluminum Powerbooks which had very minimal ornamentation that Jony’s designs started to blend form and function.

Designing with Intent is probably really really hard. Even for Apple and Jony Ive.

What Are “Services”

There is a lot of discussion on the Internet about how “services” are essential to tech companies.

Ben Bajarin recently raised the point that even though Google is using their services as a weapon to fend off the proliferation of AOSP (Android Open Source Project) devices, Google’s services are actually only relevant in markets like the US and UK, but much less so in other regions.

What you see with regard to the Google Play services availability is the biggest issue facing Google. It is one that is forcing, in a good way, local companies in those regions to create and bring to market services of their own to support their region. China is the best example of this do date. Granted China’s Android ecosystem is a bit messy with over 100 different app stores but the region is quickly fixing these issues and consolidating.

The fact that Android is being used as an open source platform is not necessarily a bad thing for Google. What is challenging is that they are not making the impact with their services the way they need to be in many of these regions. Their competition in this case is not from the likes of Apple or Microsoft necessarily but from savvy startups looking to solve a problem in their region and doing it better than Google can thus keeping Google out of regions they may wish to compete.

I totally agree with Ben’s argument, but I would also suggest that what we are simply calling “services” should be broken down into certain sub-categories. For example, looking at the Wikipedia table on Google Play availability, we see that “paid apps and games” are available in the majority of countries, whereas books, movies and music are not. Compared to the same chart for Apple’s iTunes store,, Google Play is extremely lacking in books, movies and music but not very different in apps. This suggests that digital distribution of apps is a very different business compared to that of books, movies and music.

The reason why there is a large difference is rather obvious. In the case of apps, Apple and Google are the gatekeepers. They do not have to negotiate with the content owners over whether they can distribute the content in a certain country and at what prices. They make the decisions or the developers make the decision when they submit the app.

For books, movies and music, the rights to distribute content are much more complicated. The content owners have much stronger bargaining power and they often have different agreements in each country. Each country may have their own distributor network which may have exclusive rights for distributing content in that country. Furthermore these distributors might have plans for their own digital distribution which would compete with what Google and Apple are planning to offer.

Hence the difference between Apple and Google Play is most likely the difference in negotiating power, skill and previous relationships with the content owners. Essentially, it boils down to the ability to make deals.

With this in mind, I propose that we break down “services” into the following;

self-owned services: These are the services where the provider has ownership of the content. Examples are search, social network services and web-based services (Google Apps, etc.).
self-controlled services: These are the services where the provider can distribute without negotiating with a strong content owner. The prime example is apps. App vendors are generally quite small and have little bargaining power relative to the service provider.
third-party owned services: These are the services where you are selling content that is owned by a third-party, and that third-party has strong negotiating power over distribution (unlike in the case of apps). Examples are music, books, movies, etc. Distribution of this content was historically done physically through retail networks and this resulted in complex networks and agreements, which are often different in each country. Also this content tends to be much more expensive to create than “self-owned service” content, thus requiring large companies to fund production. These large companies obviously have strong negotiating power.

When we map companies like Google, Apple, Amazon, Twitter, Facebook, Spotify, and Pandora to these categories, we find that no company is strong in all three. Twitter and Facebook are exclusively in the “self-owned services”. Amazon, Spotify and Pandora are exclusively in the “third-party owned services”. Google is mostly in the “self-owned services” and to some extent in the “self-controlled services”. They are however very weak in “third-party owned services”. Apple is strong in “third-partly owned services” and strong in “self-controlled services”. They are however weak in “self-owned services”.

From an international perspective, “self-owned services” and “self-controlled services” are relatively easy for the service provider to provide in many different countries. However, “third-party owned services” are very difficult. Amazon for example has very limited international reach. The fact that Apple has in fact been able to provide their services in a large number of countries is very much the exception.

These three categories will probably have very different dynamics and I sense that it will be very difficult for any single company to excel in all of them. At least that seems to be the case so far.