The significance of Google’s Android
22,530 views
+2 rating
Loading ...- 21 comments
- view license
Google is not an internet search firm. It’s a broker of advertising inventory.
Google makes money by building inventory (i.e. white space on web, print and radio) and auctioning off inventory to advertisers. Search is merely the means to create a boundless amount of inventory and attract billions of eyeballs to it. All Google products including Docs, Maps, iGoogle, Gmail, GTalk and News Alerts are strategies to increase the amount of inventory and attract more eyeballs.
The Android operating system for mobile phones is no different. It’s a platform for building and channeling inventory, much like a web browser. In fact we could say that Android is similar to a browser on steroids, in that it allows developers to easily build any connected handset application anywhere within the mobile user journey, and within those create more inventory.
So why is Google spending more than 200 man years building a complete operating system, instead of building just a browser for mobile phones or even a downloadable application, like an on-device portal ? Because browsers on mobile handsets are used for a tiny percentage of the time, probably less than 5% of the time the user spends on their phone. 95% or more of the user journey is taken up by the contacts application, idle screen, main menu, calendar, inbox and settings. In parallel, with Android, Google is addressing the need of handset manufacturer for an operating system they can control (it’s licensed under APL2), that’s low-cost (it’s free), that reduces time to market for variants (see the declarative XML UI framework and developer platform). Plus, Android is backed by Google, a heavyweight vendor who can support OEMs during launch of multi-million units.
What’s so special about Android ?
Android is different to other OSes, including Windows Mobile, Symbian/S60/UIQ, the Linux variants and proprietary OSes (Nucleus, EMP, BREW, etc) in several ways:
- The declarative XML UI framework enables developers and handset manufacturers to rapidly develop the user interface for new applications.
- The Android SDK is an environment for building connected applications. Every application (including dialler, idle screen, SMS, contacts, etc) can consume and produce content. Every application on Android is a Web 2.0 citizen.
- The Android source code will be licensed under the Apache 2.0 license, a non-copyleft license which allows handset manufacturers to modify the source code without being forced to share back their modifications. This is in complete contrast to GPL v2 and GPL v3 which is a copyleft license (see our white paper); Sun applied the GPLv2 license to its Java ME implementation, which is the reason why not a single handset OEM is using it.
- Android allows developers to program against the familiar Java SE library of APIs (the desktop version of the Java libraries), which is much broader and more powerful than Java ME, the mobile version. Much like SavaJe (now Sun’s Java FX Mobile), Android is a Java SE -like platform built on a Linux kernel, but more importantly one where the Java platform is deeply integrated with the underlying Linux support package. In other words, the Java SE-like platform is a native application platform for Android phones. Symbian may arrogantly dismiss Android as yet another Linux initiative, but the breadth and depth of Java APIs is something Symbian never managed to get right. And unlike the FX Mobile platform, Android has several OEMs who are planning to build handsets on it.
- Android is not only a departure from Java ME development model, but also away from Linux development. Funnily enough, operators like Vodafone and Telefonica who have committed to supporting Linux as a prefered platform would not be counting Android in. (thanks Guy!).
- Android uses Dalvik, a ‘proprietary’ (non-Sun-endorsed) Java virtual machine which means that Android developers can use Java SE APIs, while Google does not have to pay any royalties to Sun for TCK certification, as they ‘re not claiming this is a Java environment. As Stefano writes, Google doesn’t claim that Android is a Java platform, although it can run some programs written with the Java language and against some derived version of the Java class library. This is slap in the face of Sun.
- Google is paying developers $10 million to write applications for Android, which is a smart move to motivate developers especially when no phones are out yet. It’s worth noting that $10 million exceeds the yearly marketing budget of most operating system vendors.
The Open Handset Alliance (OHA) is formed by an array of complimentary participants; operators (covering US, Europe, Asia, Japan and Latin America), handset OEMs covering all global regions (including HTC, the second-biggest smartphone OEM after Nokia), as well as hardware and software vendors covering complimentary constituents of a mobile handset.
So is Android mature and will it be adopted by OEMs ?
Google has dedicated an estimated 200+ man years building the platform (since the Android acquisition), but there are still bugs (see this report). HTC has confirmed it is launching one handset in 2H08 and reportedly plans to release a total of 2 or 3 Android-based handsets in 2008. Moreover, according to a WSJ report, T-Mobile US has committed to releasing a phone in 2008 that will be based on Android. For a new Linux initiative, this level of commercial support is extremely rare.
What’s in it for Google ?
Android is a service access platform, not a delivery platform. It’s about growing the pie of mobile advertising inventory and not necessarily growing Google’s share.
There’s nothing to stop Yahoo taking Android and launching a phone with Motorola that bundles Yahoo Go!, flickr and eBay. I ‘m guessing however that Google has some sort of agreement with OHA-participant handset manufacturers and operators about bundling Google services with Android handsets by default.
Moreover, Google might want to bundle the gPay payment system (see this Times Online article). Or connect the physical world to Google advertisers via its ZebraCrossing QR reader technology for mobile phones.
What’s even more interesting is that it may provide a channel for feeding customer analytics back to Google, such as presence, contacts, call logs, SMS messages and a wealth of user profile information that can be used to build extremely detailed digital footprints.
Another important impact of Android is that it will catalyse the development of white-label phones, i.e. phones ready-to-customise by consumer brands like MTV, Nike, Gucci and Tag Heuer. Rapid software customisation is what hampers the scalability of customised design manufacturers like ModeLabs today.
All-in-all, Android seems to be the only non-proprietary operating system with a strong chance of wider commercial adoption. Motorola is losing interest in LiMo (it committed to Qtopia APIs, whereas LiMo supports rival GTK). The LiPS forum doesn’t really have a route to market, apart from Chinese ODMs, and is a partial OS. All other mobile Linux operating systems are either in alpha stage (Celunite, ALP, A la Mobile), not shrink-wrapped (Greensuite), or not backed by a big services firm (Purple Labs). Symbian is dominated by Nokia and DoCoMo; outside Japan, the overwhelming majority (volume-wise and model-wise) of Symbian handsets are Nokia, whereas in Japan the vast majority of 30 million Symbian-based shipments are DoCoMo (60 out of 66 models). And Windows Mobile is for enterprise segments only (at least up to version 6). Plus Android ticks several boxes of OEM checklists including control, time-to-market and cost.
Thoughts ?
- Andreas
Very insightful view point, and interesting. My view is that this is a realisation that the mobile phone will form the basis of the next computing platform – and Google are trying to make a grab for this quickly, by creating the service access platform you describe. Posted here if you’re intersted – http://jonathanmulholland.com/2007/11/05/personal-computing-will-be-mobile-what-microsoft-did-to-the-pc-google-plans-for-the-smart-phone/
What about Nokia’s 770/N800/N810 line of open source Internet Tablets based on Maemo Linux? How do they relate to Android?
Jadon,
Maemo is the open source application environment which runs on top of Nokia’s Linux distribution for Tablets. It is not shipping on any phones with cellular capability, only WiFi-connected tablets.
As an app environment, Maemo competes with Android’s Java SE-based app environment.
Andreas
Another possibility. Google originally wanted to develop its own branded phone when acquired Android. Later on it realized a couple of things:
1. It takes too much effort to do a full phone than initially thought
2. Even if gphone pulls off, the market share is at question. Without enough market share, it looses the whole purpose.
So they decide to open source it and give it a shot.
Technically I don’t see any reason why Android would be successful. However given the name and backing of Google, anything is possible.
No. Google is first and foremost a search firm — good results quickly. As soon as the results suck, or a couple of PhDs come up with a much better search algorithm the rest goes away (OK… not goes away, Google just remains another really big company, but nolonger the best and brightest.)
Andrea,
What would the benefit be for an operator to launch handsets with the Android platform?
Giorgos.
George – the benefit would be a share of ad revenue (as per Vodafone’s and Telefonica’s deal with Nokia’s Ovi). Android’s special in that you have a much broader inventory ’surface’ across the user journey compared to other ad channels which only use SMS, video, browser, games and idle screen. Broader inventory surface implies more ad revenue, and more ways of selling and cross-selling content, too, for operators.
Andreas
Fascinating article. Certainly Android, despite the bad name, will have an effect, the question is the extent of the effect. Open source software only takes off when a large number of people participate in the core structure while all other enthusiasts/small developers/hobbyist fill in the gap for auxiliary software that works based on the platform. As for Android, it is yet to be seen how the core is developed and how well maintained it is. The combination of Java/Linux only appeals to certain number of developers and not everyone may jump in. As it is well known, there is tremendous amount of code reuse and architecture reuse going on these days, so a whole new system may require a significant amount of effort and commitment.
All in all, its a good thing that Google has bothered, but it may be slightly late, or slightly out of sync with those who develop software on mobile devices. We shall see how it goes …
This person has shamelessly lifted your article
http://designtraffic.blogspot.com/2007/12/what-is-so-special-about-googles.html
Thanks Mahesh.. very kind of you to let me know. I have left him 2 comments on his blog.
Andreas
Great article. Thank you.
Javier Marti
Trendirama.com
Hello, Yah has found many links to those question and has published it in a Spot.
http://www.jamespot.com/s/344-Google-Android.html
The article was extremely informational, but didn’t take much of a stand. According to the article, Android has “a strong change of wider commercial adoption.” But the article also makes it clear that anyone–like yahoo–can bundle up Android and customize it with its own mobile apps and strike a deal with the carriers and OEMs to use the Yahoo version of the platform.
This is good and bad for Google. They obviously chose to go down this risky path for a reason. This reason is that it will allow for wider adoption in handsets. The risky side for Google is that other companies will be able to harness the technology without paying tribute to google (paying tribute would be in the form of services monetized by google ads).
It therefore seems Google is relying on loyalty, trust and other intangible emotional/human motives on the part of the Carriers and OEMs. It’s like they think: because we are the ones that developed this and put all our energy into it, the OEMs and carriers will want work with google to monetize the platform.
Earlier the example of Yahoo using the Apache 2.0 licensed Android platform was used. But a better example is if the OEMs and carriers do it themselves, monetizing the platform without any help from the software giants.
…So I can’t believe Google would be this stupid to just GIVE us God–um Google’s–gift to mobile phones. They must have something up their sleeve. Maybe it’s as simple as they already struck all the deals with the OEMs and carriers. Maybe it’s more deceptive: something kind of like what Apple has done that makes the Safari browser faster than the FireFox 3.0 browser in that Safari has access to special Apple development APIs that FireFox does not. But I doubt this, as it would destroy Google’s “do no evil” image.
So in conclusion they’re probably just going to out-develop everyone with their platform. Android is probably purposely built so that it is hard to force users to use the exact applications that come with it. It’s probably very dynamic in its ability for users to add and delete applications–it’s not going to be like Windows which gets most people using Internet Explorer as the web browser. So for the non-google endorsed Android platforms, Google probably has a plan to make users want to replace the default installed apps with google tools!
That’s all I can come with for now. Any other guesses?
In response to the previous comment:
The controls that Google will add to discourage non-Google-enabled versions of Android is somewhat of a puzzle..
The two things Google might have up its sleeve is:
1. Keep some part of Android as closed source (it never said that 100% will be open). In fact the Android SDK isn’t yet under an Apache license.
2. Make it exceptionally easy for developers to develop apps based on Google back-end services (gmail, google apps, etc).
I would guess both above means would be used in the final Android SDK.
- Andreas
Worth reading this article, What an idea of doing the business…
I would like to know what’s the difference between the Nokia SDK from FORUM NOKIA and Android SDK.
Can some one can help me out knowing the difference
I would like to know what’s the difference between the Nokia SDK from FORUM NOKIA and Android SDK.
Swathi – Nokia SDK is for developing apps on S60 phones (100 million+ of them). Android SDK is for developing apps on phones that will ship with the Android stack (0 of them and counting..)
Andreas
From a lay-person’s point-of-view, how would a developer, even a newbie developer, go about monetizing their app on the android platform?
Aside: I reckon that Google has gone about building an ad-brokering empire far more subtly than the likes of Microsoft built their (multiple) footprint(s). Somewhere at its startup, the people at Google and their advisors realised that getting and retaining so many eyeballs would lead to them cornering at least some forms of the online ad market. The core Google search algorithms and concurrent use of the low-cost, high-availability hardware/software/connectivity search-serving platform was/is foremost in their arsenal, but I’d say that the ability of the masses “to Google something” (How many people use the terms “Go Yahoo it” or “Go Lycos it”?) is simply a means to an end, and that end is the formation of a very large ad media serving/brokering footprint. I’m no fundi on keeping up to date with their share price trends, but I’d agree that going the route of developing android is a move to make the the online ad pie bigger and so sustain their increasing market cap via increased real/perceived future ad revenues. Perhaps when “TV” integrates IP on the same screen like a mashup of web and TV on one screen in your living room, they’ll have succeeded in making the pie even bigger?
Can you give out your comments on Symbian and Nokia? We know that Moto will work with Nokia.
Hi Jay,
Have a look at the full analysis of facts and repercussions of the Symbian Foundation here.
Andreas
visionmobile 2005-2010
visionmobile 2005-2010