Finding the Oasis in the Android Desert

It’s tough to make money in the Android universe. When I first started reading about Android smartphone manufacturer profits, I was shocked by how little, if any, profits were made selling Android devices.

How could Android smartphone manufacturers thrive in this environment?


Android is open source, meaning it’s available under a license to let anyone study, develop on top of it, and distribute it. It’s conveniently called the Android Open Source Project (AOSP). If you’re a smartphone manufacturer (from here on out, will refer to them as Original Equipment Manufacturers or OEMs), you can build off of the AOSP for your devices, by agreeing to the “Android Compatibility Definition Document” that forces a standard, so that apps and services are guaranteed to work somewhat consistently across devices [1].

However, if an OEM wants access to Google apps, they have another agreement to sign with Google. This further restricts how differentiated an OEM can make their software experience.

The catch is: an OEM can’t pick and choose from the Google apps. It’s an all or nothing deal. If they want anything Google, they have to pre-install list of Google apps on their device, and only use the Play Store as the distribution channel for apps. And if Google doesn’t like any of the software modifications an OEM makes to their version, they can prevent access to their apps and services.

Google apps are tremendously popular [2]. This puts pressure on OEMs to comply, and provide access to these Google apps on the phones that they sell. Google is the house in this game: they capture most of the value. All apps are downloaded through the Google Play Store. Every time someone makes a purchase for an app, or within an app, a percentage of that transaction goes to Google.

It’s Difficult for Manufacturers to Make Money

Margins only really exist at the premium end of the smartphone price spectrum (think iPhone, Samsung Galaxies and flagship phones). Manufacturers realize this, so they release their latest and greatest “flagship” each year, priced at the top of the bracket.

This makes this segment hyper competitive. Due to the higher margins, there exist opportunities for OEMs that can wait out short term profits for a larger piece of the pie (e.g. OnePlus, a Chinese OEM, undercuts flagship phones, offering nearly the same specs as flagships, at a significantly lower price point).

Competing on hardware specs alone isn’t defensible. We’ve seen this with the churn in OEM standings in the last 5 years. When an OEM offers nothing more than a laundry list of specs, they are easily overthrown by the next OEM that figures out how to add more specs to the list / offer the same specs at a lower price.

How do OEMs Differentiate Themselves

We’ve reached an inflection point — smartphones are powerful enough to handle computationally intensive experiences like virtual reality. OEMs are in a powerful position when it comes to creating these experiences because they can go below the application level and add functionality that 3rd party app and peripheral developers can’t (without branching off and making their own versions of Android).

For example, some things that only OEMs or forks of Android could have pulled off as a result of their ability to make changes lower in the stack:

  • Cyanogenmod’s (a discontinued OS based on Android) multi-window functionality, could have only elegantly been done below the application layer (in the framework layer where the logic for how apps are rendered on screen is located). This feature allows two apps to share the screen. Google incorporated this feature into Nougat (the latest version of Android), but before this update OEMs (or OSs branched off of Android) had to implement it themselves if they wanted to provide this feature. [3]

  • The “OK Google” assistant’s “Always On” feature (where you can say “OK google” even when your phone is in standby, and it wakes up to assist you), is only implemented well on a few devices. To implement this feature well, OEMs need to make modifications at the hardware and firmware level of the stack.

  • Samsung’s virtual reality headgear, GearVR, only works on Samsung devices with optimizations such as reduced latency (which is a crucial factor affecting usability), which would be difficult to provide without making the modifications at the firmware, and hardware level of their phones.

  • Samsung’s Security layer (KNOX): Samsung wanted to target enterprise before Google’s Android was ready for it. So they introduced their own security layer at the system level that gave organizations the ability to efficiently, and securely manage their employees’ devices.

A Smartphone ecosystem that offers users a unique experience and optimizes inter-operability between its devices, is more defensible than a laundry-list of specs. OEMs realize this and are making efforts to connect various devices: VR, Cars, bikes, POS systems to offer that differentiated, seamless experience.


Don’t forget who the house is, if you rely on Google’s distribution, apps, and services, you’re building your castle on their land. And Google isn’t necessarily happy with some of the innovations that OEMs have come out with.

For example, Google put a lot of pressure on Samsung and Cyanogenmod against releasing their multi-window features. Some of Google’s concerns were well founded: if an app didn’t work well in this new mode, it would result in a hampered experience, creating a scenario in which there would be ambiguity over whose fault it was — the app developer, OEM, or Android itself (and therefore Google by association).

One could argue that if an OEM wants to truly differentiate itself, it needs to take control of its own destiny by breaking away from Google and its apps and services. A few companies had the courage to do so, but…it didn’t go as well as they’d planned (e.g. Amazon’s Fire OS).

Because it’s damn tough. Most companies aren’t capable of building a world-class suite of apps to compete with Google’s offerings: Chrome, Maps, Docs, Youtube to name a few.

With more and more players entering the smartphone manufacturing business, the pressure to differentiate is growing. I predict we’ll see more companies attempt to break free from Google’s shackles, out of the desire to innovate (read: survive).


[1] Android developers already deal with the headache of fragmentation — which is that the latest Android version is adopted at a slow rate. Which means developers have to develop for multiple versions of Android, adding to the development time of their software. The more the inconsistencies between OEM implementations of Android, the greater the cost of developing an Android app. It’s in everyone’s best interest to make it easier for developers to build apps for the ecosystem.

[2] Google apps are extremely popular… outside of China. China is a completely different ballgame that we’ll cover in a later post.

[3] We’re also building a multi-window experience that resembles what you see in a desktop OS: ability to resize, minimize, maximize windows for our product, the superbook. It’s tricky to say the least. If you’re interested in transforming smartphones into laptops, reach out to us! We’re actively looking to hire talented Android engineers.



Thanks Luke XieAllison Donine, and Jesse Sum for reading drafts, and providing valuable feedback.

I originally posted this on Medium.