Predicting Google’s Phone Ambitions

This is a fairly long post. To make it easier reading, it has been divided into three parts

I.) Introduction: This highlights the existing motivations for Google to build its own phone

II.) Predictions for Google’s Phone Moves: This summarizes the entire post while making some obvious and some less obvious points about Google’s phone plans. This is a logical place to pause reading for people who want to skim the post.

III.) A Detailed Analysis: This contains the approach of how we reached our predictions while analysing competitors and the Phone industry.

Finally, this post is restricted to Google phones (Pixel) and not a larger Pixel platform that Google probably has in its radar involving Tablets, Chrome OS-Android integration & smaller devices etc.

I.) Introduction

Google recently announced their Pixel mobile phones which have been getting rave reviews.

These are the first mobile devices which are Google branded and go beyond what Google has attempted so far -:

• The first move was an acquisition of Motorola which was subsequently sold off to Lenovo. The motivations for the entire buy-sell episode have been analysed in many different dimensions, but the bottom line is Motorola under Google was not manufacturing phones for too long
• The other attempt was launching the “Nexus” brand of phones having stock / near-stock Android but being manufactured and sold by other OEMs. The aim here was to demonstrate flagship-grade Android phones leveraging the best of Android’s capabilities

Reasons for Building a Phone

The attempt to “build” (“build” has various meanings as shall be seen throughout the article) its own phone is closer to the first step above. The professed reasons for this are many -:

• Showcase consistency in Android: The most obvious one is akin to the attempt with Nexus i.e. to showcase the best of Android. Unlike iOS, Android is highly fragmented and upgrading the software version is highly dependent on the OEM. This causes different behaviour for developers & consequently for consumers. Bigger issues are related to bug-fixes and software optimizations which can drastically alter the Android experience.
• Showcase best of Android: The other part is to do with showcasing Android as the most premium experience possible. The best software is only as good as the platform it runs on. An underpowered device may not be the best phone to capture great media today or showcase VR in future. Hence the attempt to build flagship phones instead of focusing on the entire spectrum of price range. Focusing on high end helps Google worry less about marketing at price points and more about features. Also it is known that benefits of products at high end of pricing eventually start trickling into the lower end of the spectrum
• Showcase best of Google : There is an entire suite of Google applications which are mandated to be pre-installed by OEMs when using Android. Applications like Google Maps become more useful and contextual with increase in number of installations. But the real meat is in Google as a data engine beyond search. As it keeps collecting data from mobile devices which closely mirror our personal interactions, solutions like Google Assistant can be entrenched deeply into the devices. This will soon lead to the Operating System overriding the Mobile app space for a lot of non-gaming use cases. Once it starts working as well as it is supposed to (and it already works decently), Google Assistant is the one app to rule them all.

I would perhaps even argue, if not for the potential of a new form of data through personal assistants like Google Assistant, the motivation of a Google building a phone on its own would be much lesser.

Biggest Reason for Building a Phone: It’s the new PC

Google would like to showcase the best possible Phone because it is the new “PC” - a Personal Communication device.

Put differently, a phone is an invaluable device because it is an extension of personal behaviour as well as a representation of a node in multiple social networks. It represents both, the convergence of and divergence of an individual & the society.

Phones enable communication of different forms but more importantly the communication channel is no longer a passive medium.

1. The participants in a communication involve people, devices, information, data and actions
2. The channels are listening to data and processing them to continuously influence and shape communication

The level of insights possible at a micro-level will make the winner of a phone channel extremely powerful and will be used to extend presence to other devices as well

At the end of it, Google with a DNA of managing high volumes of data, is entering the Phone market, because it wants to build on its advantage of being able to collect, manage and present the most contextual data in the fastest possible manner

II.) Predictions for Google’s Phone Ambitions

Before we enter into the details, we have listed some of our predictions related to Google Phone-:

1. Google will enter Chip-design and offset the costs involved across various devices not restricted to Phones
2. Google will like to keep the Android industry hyper-competitive. This will enable it to get manufacturing of SoC by competent vendors who lost the mobile processor race & manufacturing of phones by those who lost the Android Phones race
3. Google will want to open up Chipset design (currently considered as a competitive advantage) to outflank competitors and keep costs on entry into the Android ecosystem low
4. Google will then be in a unique position of being a vertically-integrated open-sourced player with Chipset (like ARM) — OS (Android) — Apps (By Google)
5. Pixel & Android will be independent units in Google (or more correctly Alphabet) with Android treating Pixel at par with all other OEMs
6. Pixel will need to show OEMs the way to build process agility in working with Android. This might lead to two types of devices from the OEMs- one with an Android version which is at par with Pixel and the other flagship device with some differentiation where the OEM’s signature is more pronounced but (probably) a lower version of Android
7. Operating System (with interfaces like Google Assistant) will increasingly eat into Mobile apps space and non-gaming based applications will find a hard time to survive
8. A new type of “application” called “installed Websites” is proposed which would be Webservices from application providers which can be “installed-to-be-invoked” from the Operating System’s interface

Now that we have summarised our predictions, a detailed analysis follows for our predictions

III.) An Analysis for predicting Google’s next moves

Given Google’s financial strength and its existing presence in the Mobile landscape, one might consider success in building a phone to be a given. There are some major challenges though, both within and outside Google which will need to be resolved to achieve any semblance of success -:

1. Level of Vertical Integration
2. Android’s Relationship with Phone OEMs
3. Google’s Relationship with Phone OEMs

Let us look at each of these to understand the complications and potential threats to this decision made by Google

1. Level of Vertical Integration

With the omnipresence of Apple’s success these days, it is often easy to mention that whatever they are working on is the best way forward. So vertical integration gets spouted freely. But companies like RIM, Nokia were vertically integrated to a large extent. We need to go one level deeper to understand vertical integration and its potential impact.

Deciding “What” to Integrate

From a historical standpoint, the Computing Industry was always driven by Vertical Integration till as recent as the 1980s (and this is still true in larger systems. Just in case you did not know, vertically integrated mainframes are still around and very much in use). The major companies prior to the PC revolution, like IBM & DEC, were more vertically integrated than the Apple of today. These companies would do everything, chip design, manufacture them, the OS, build the servers. On the other hand, in Apple, which is considered as an example of vertical integration, the latest processor in iPhone 7 is designed by Apple but manufactured by another party , the physical camera, the retina display are other examples where Apple is not involved in the manufacturing.

To decide what can be vertically integrated, perhaps the existing template provided by Apple is a good one to start from. So it follows that Chip design is a good place to start. But there is a catch -

Even Apple does not follow Apple — the Macs have been run by processors designed by Intel since 2005 but the iPhone has chips designed by Apple since 2010

So clearly the “What” is not as easy as answer inspite of there existing an existing successful model of Apple. There are various nuances in deciding the “what to integrate”. Most components rule themselves out as not being important enough. But there are two key areas where Apple focuses on which are lacked by others -:

1. We could start from the most obvious candidate, chip design or more specifically the System-on-Chip (SoC). The SoC plays a significant role in most key areas and Apple’s GPU plays a huge role in differentiating it from the rest. But to reach where it currently is, has taken it years — from 2007 (first iPhone) to 2010, Apple was essentially working around ARM architecture (even the A4 was a modification of ARM design). Building skills around this area takes a significant amount of time. Even more difficult though is the fact that chip design is very expensive, only the volumes of iPhone and allied devices like iPad justify the spend on internal in-house chip design.
2. Fabrication capabilities is the other differentiation which has been built by Apple over the years. While their primary focus has been building expertise in aluminum , they also have a known history of playing with materials with Ceramic version of Apple Watch being the latest example. This aligns well with the Design-first approach which Apple is renowned for and is perhaps something which Google can & will pay less emphasis to.

Building capabilities around fabrication of new material may not be a priority in Google’s radar. However the huge benefits of Chip design make it an obvious candidate for Google for integration.

Chip design is turning out to be so critical, that companies like Huawei & Samsung are already have started moving in that direction

However, building Chip design skills requires significant time, it is probably a given that Google is working with partners around chipsets in the near future.

For the long-term though, Google is definitely interested in building inhouse expertise in SoC but their challenge is that justifying such a spend will need Google to achieve sales volume equivalent to the Apple stable. It will be difficult for a Pixel phone alone to reach it so most probably, the idea is to include a stable of Pixel products which can leverage the SoC across multiple product lines. But they have one major trick up their sleeve which is unique to the Android ecosystem they play in:

The biggest power Google has over Apple is to treat their own Chipset design at par with their philosophy with Android i.e. publish the Chip Design openly like ARM. This has HUGE implications in spawning an entire ecosystem of new devices powered by Google’s Chipsets (on top of existing OS and software)

This route has surprisingly few companies as competition; a lot of incumbents in the CPU and GPU markets have been struggling for years now . Unlike earlier when Intel refused to work with Apple for its iPhone, there will be many companies today willing to cooperate with Google in the Chip Design space else a lot of them might risk becoming irrelevant in the market.

One final note — Google can take the route of Apple and avoid mass-manufacturing or take the Samsung route and build capabilities in the manufacturing space. It is a given for Google, that even if Chip design is done inhouse, the actual manufacturing of it will be done by a 3rd party supplier. Managing such suppliers is the next challenge of vertical integration

“Managing” How to integrate

The sheer volume & diversity of components packed in a phone mandates that a variety of suppliers are involved with any phone release (we saw above that even the most critical of parts like SoC are manufactured by outside contractors). Apple has over 200 suppliers to manage .

This implies that the integration touchpoints, the quality, the inventory needed across all suppliers needs to align with the larger goal of product release cycles. The ability to execute this at scale is one of the most important pieces of vertical integration

1. Not only does a company have to deal with suppliers, it has to deal with relationships between suppliers
2. At the same time, the company still needs to ensure enough isolation between suppliers to avoid getting caught in a hostage situation with a few key component suppliers
3. It also needs the ability to quickly switch suppliers incase of non-performance — Apple has done this multiple times as a business (e.g. PortalPlayer to Samsung for the iPod) and this requires a high level of niche technical skills (move from PowerPC to Intel for the Mac), and not just vendor management techniques
4. It involves working with suppliers which are often competition —for iPhone, Apple works with other major phone manufacturers like Samsung, Sony, Lg and the case with Google’s Pixel is no different.
5. The prices negotiated with the suppliers need to be low and volumes of a new Phone by Google may not be enough to get significant leverage

Building a highly efficient supply chain for vertical integration is a time-consuming process and Google will probably not want to focus in this area too much. With the initial number of SKUs for Pixel Phone professed to be small combined with the fact that none of Google’s existing hardware initiatives have helped build a competitive supply chain, this is an area which Google will probably keep an eye on, but will not over-invest.

This hyper-competitiveness of Android phones will actually help Google on the manufacturing side as well. For every Huawei which was not keen on a Google-branded phone, there exists an HTC which needs to find its place in the market & hence is eager to help Google

Conclusions of Vertical Integration

The high level of adoption of Android, the sheer number of companies which lost the microprocessor & mobile phone wars and Google’s deep pockets will give rise to a unique Vertical integration model where -:

1. Manufacturing of SoC will be done by highly competent vendors in the Processor space who have fallen behind due to the prevalence of Mobile phones and the associated chips

2. Manufacturing of phones will be done by highly competent vendors in the Mobile space who lost the Android wars

3. Chipset Design will be opened up by Google to create a vertically-integrated model of Chipset (like ARM) — OS (Android) — Apps (By Google)

2. Android’s Relationship with Phone OEMs

OEMs adoption of Android was swift and it enabled them to focus on non-OS related aspects of the phone. Starting a new Billion dollar phone manufacturing company is easier because of Android — players like Xiaomi, Oppo, Vivo from China; Micromax, Lava, Karbonn in India have all grown rapidly using Android as their backbone.

Android cannot be replaced by OEMs

The hyper-competitiveness of Android made a lot of larger incumbents jittery with Android and many have tried to get out of its hold.Various strategies were adopted in replacing it but none of them could fly :

1. Incumbents Failed: Microsoft/Nokia, Blackberry could not create an ecosystem as Android/iOS had moved too far along. And this is inspite of Nokia’s excellent distribution capabilities and huge mindshare in growing markets like India
2. Content-providers failed: Amazon’s plan of forking Android and creating a new phone did not fly inspite of having a significant amount of content available with them
3. A Vertically-integrated Approach failed: Samsung’s plan of moving towards Tizen gradually from Android was nipped at the bud by Google
4. Niche players still finding their niche: There always will be players in the market trying to find their niche but these need to be large enough to be considered as a major player — Cyanogen & Sailfish being a couple of examples. A notable failure in this category is Firefox OS which I think was an idea too early for its time.

Google’s well-populated Play Store was the number one reason that these initiatives were late to the party.

OEMs cannot differentiate in Application Layer

The way it worked in the Android world was that OEMs would take Android and plug in a skin above it to create a customized feel specific to their brand. They also created their own set of applications (and even application store). This approach has led to the heavy fragmentation of Android, bloatware and the consequent experience issues with Android.

Google’s array of applications like Google Maps which are extremely amenable to mobile usage, desktop-based alternatives of Mail, Search, Browser; cosmetic applications to manage media, news etc. backed by ample storage and the fact that it is mandatory to install applications from Google for any OEM licensing Android, made the entire application layer in Android extremely sticky.

This coupled with the fact that new software upgrades would become easier, forced OEMs to move away from heavy customization of Android with the maximum probably being retained only by Samsung and that too at a level which is comfortable for Google

Pixel = (Just Another) OEM for Android

With little scope to play with in the OS & the application layer, as well as the fact that chipsets used by OEMs are common, the scope for differentiation is in the hardware, system software for tasks like image processing and form-factor.

With limited scope to differentiate with, no OEM can feel that a Pixel device receives benefits of a different version of Android. So even though OEMs need Android, it is in Google’s best interests to play with a straight bat with Android.

Conclusions of Android & OEMs

This implies the following -:

1. Pixel & Android will be independent units in Google (or more correctly Alphabet)

2. Android will treat Pixel at par with all other OEMs providing preview/access to its builds at the same time across the board

3. To manage the process, Android will need to modify their Release Management process to make it easier for OEMs to integrate and upgrade Android. This also implies that Android development will need to be a couple of cycles ahead of the version released for OEMs (including Pixel)

4. If Pixel can build an agility in working with Android, this will benefit the entire ecosystem as these lessons will be imbibed by the other OEMs as well. This might lead to two types of devices from the OEMs- one with an Android version which is at par with Pixel and the other flagship device with some differentiation where the OEM’s signature is more pronounced but (probably) a different version of Android.

3. Google’s Relationship with Phone OEMs

There is an entire suite of Google products in Android OS which are outside the purview of the Android team. There are two types of these apps -:

1. Apps residing in the Application layer:

These include applications like GMail, Maps etc. which are mandatory be be installed by OEMs as part of their OEM agreements. Some of these like the Play Store, Google Search & Maps are probably better than anything the OEMs can come up with by virtue of the sheer data and users these applications have. Others like Google Play Music typically had alternate options from OEMs.

2. Apps residing deeper than the Application Layer:

This is where things start getting trickier for Google. Building a product like Google Assistant requires a deep level of integration with the operating system. Apps like this need to be integrated or deep linked into other applications within Android. Infact, it is possible, that most use cases can be killed if the information provided and the actions taken via applications can be exposed via the Google Assistant interface.

The best possible place to develop such products is within Android. However, Google Assistant is a more horizontal offering which needs to also reside in as many consumer-facing devices as possible, the Google Home running on Chromecast being one such example .

Infact, the aim for Google would be to run Assistant across as many devices as possible agnostic to the underlying device or OS. This implies the service will need to be a horizontal offering with different flavours based on the end device it is running on

The challenge here for Google is to decide whether it wants to achieve parity of Assistant features across all Android devices (which is a larger set of data to collect from) or have a version in Pixel devices which is slightly ahead in terms of capabilities. Google will want the former (for sake of data) but will find it easier to execute the latter.

Conclusions of Google & OEMs

1. The way out would be to continue with deep-linking and making changes in Android which expose data from installed apps via APIs which anyone can use.

2. The other piece which could be implemented is exposing data of not just installed apps but of “installed Websites” which would be Webservices which can be “installed-to-be-invoked” from the phone

The Operating System wars are over with the duopoly of iOS and Android well entrenched. It is now the time for wars & of consolidation in the Phone evolution space….