This guest post is written by Vic Gundotra, Vice President of Engineering for Google’s mobile and developer products. (Prior to Google, he spent 15 years at Microsoft, most recently as their GM of Platform Evangelism.) Vic credits his now-7-year-old with forecasting the importance of mobile data access, and now carries at least 4 phones at all times. Fortunately, he had two kids before adopting the possibly-prophylactic habit.

---

Focus on the mobile user, and all else will follow

Simpler data, better browsers, and a smoother experience

Today the mobile industry finds itself in a unique position to do right by its users:

Worldwide phone penetration continues to climb at a break-neck pace, with over 4 billion mobile subscribers at last count.¹ (In comparison, the PC industry is forecasted to see its sharpest unit decline in history.²) Prevailing economic conditions will accelerate this trend, as users consolidate pricey communication services into cost-effective, all-in-one mobile devices.³ And for the first time ever, half of all new connections to the internet will come from a phone in 2009.⁴

Google’s mobile traffic reflects these milestones — having quintupled since 2007⁵ — and it underscores users’ appetite for mobile data services. But as a community of operators, device manufacturers and software providers, we continue to get in their way. In short, and as a general rule, we make it too costly, too unfamiliar, and too difficult to do anything beyond voice calls.

In reply I offer up three suggestions: simpler data plans, better web browsers, and a smoother on-device experience. And in each case I’ll use Google traffic numbers as a proxy for total internet usage and user happiness.

Disclaimer: As a Google employee using internal data to carry the weight of this article, I owe it to the reader to lay bare my economic incentives: the company I work for has a financial interest in the broad and sweeping adoption of the Internet-as-we-know-it. Indeed, more internet users leads to increased web usage, which often leads to more Google searches and downstream ad clicks. I use Google data because it’s what I know best, and because it reinforces my industry-facing remarks, but make no mistake: I’m fundamentally interested in what’s good for the mobile internet. It just so happens that this is also good for Google. With that said, I hope you’ll find value in the words and data that follow.

Flat is the new phat

Consider MetroPCS, a regional carrier in the United States with just over 5 million subscribers on their 2.5G CDMA network. Over the past year, their Google search volume grew over 2.5x more quickly than another global carrier with 10 times as many users, and a 3G network.⁶

Metro’s “secret” is a free month of web access at signup, with the option of flat-rate, unlimited data thereafter.⁷ As a result nearly half of Metro’s subscribers use the web on a regular basis. (It’s also worth mentioning that MetroPCS was recently recognized for excellence in customer satisfaction.⁸)

In contrast, many operators subject users to a labyrinthine set of data options, from pay-as-you-go to daily caps with significant overage charges. Now, can you imagine paying your at-home internet provider for every page load? Or needing to know the size of a website before visiting it? Or managing your monthly download quota across your entire household? It’s simply not practical, and it’s all the same internet, so why do we treat mobile users as second-class citizens? Case and point: my colleague’s January phone bill contained 27 pages of itemized data charges, spelled out in excruciating detail.⁹

Unless we declare flat the new phat — and soon — I fear Occam will do something terrible with his razor.

They want it all, they want it now

Users “get” the web, and they’ve known for over 10 years that the browser is the thing that takes you there. Likewise, more and more of today’s killer applications are the Amazons and Facebooks of the world, not software that you download to a local machine. So it should come as no surprise that mobile users want phones (and browsers) that put a fully-featured internet in their pocket.

For example: the availability of a modern web browser explains why iPhone and Android users — just 13% of the high-end market¹⁰ — represent nearly 50% of Google’s smartphone traffic worldwide.¹¹

Similarly, users of the T-Mobile G1 and its newer WebKit browser search Google 20 times more often than Nokia Series 60 users.¹²

Both data indicate that it’s about usage — not just units — and this trend will continue unabated with more efficient JavaScript engines, and more sophisticated HTML5-compliant browsers.

The simple truth is that mobile users have wanted fast and full web access all along. Consider two quick facts about Google search behavior: the “tail” of PC and iPhone queries is significantly longer than that of feature phone queries;

and the gap in query diversity between desktop and high-end mobile devices is shrinking.¹³ People want all the world’s information on their most personal of personal computers, and we need to offer browsers that scratch this quintessential itch.

“One web will triumph.”¹⁴ Users want all of it. And they want it now.

Friction is fugly

In the early days of mobile search, customer feedback was clear: “I can’t find Google on my phone.” And in hindsight it makes sense: unintuitive device menus and preference panes mandated 20+ mind-numbing clicks just to locate portal content¹⁵ — nevermind “off net” sites like Google. This Frankenstein’s monster of OEM, carrier, and 3rd party software made it impossible to discover — much less enjoy — mobile data services, and showed a complete disregard for users’ on-device experience.

Thanks to an influx of smarter phones, many mobile users can now reach 3rd party software with a single tap or click. And in Google’s case, this desktop-like experience increases search traffic by many orders of multitude.¹⁶ Why? Because it provides a frictionless onramp to search results. Likewise, and prior to its v5.0 release in February 2009, Google Earth saw more activations on the day of its iPhone launch than any other day in the product’s history. Why? Because the iPhone’s App Store and on-screen layout make it easy to find, try and access mobile data services.

And herein lies the rub: users appreciate well-written software, but ease of use and on-device navigability are critical preconditions for usage. (After all, if you hide a tree in a forest, who cares whether someone hears it fall? Chances are they’ll never find it anyway.) The proliferation of app stores is a positive step in this direction, as are efforts on the part of OEMs to give developers unfettered access to low-level functionality.

We have to surprise and delight users with fast and fluid interfaces. Friction is just fugly.

- Sent from my Android phone, with a WebKit browser and an unlimited data plan

---

1. ITU, 2009
2. Gartner, 2009
3. comScore, 2008
4. eMarketer, 2008 and 2009
5. Google internal
6. Google internal
7. MetroPCS, 2009
8. J.D. Power, 2008
9. January phone bill, redacted
10. Canalys, 2008
11. Google internal
12. Google internal
13. “Computers and iPhones and Mobile Phones, oh my!”, 2009
14. Opera, 2008
15. http://www.biz-lib.com/products/ZMOMX.html
16. Google internal

Android is the new highly-anticipated “open source” mobile platform that is currently under development at Google. There are a high expectations of Android as the first viable open source based mobile platform and a viable competitor to the dominance of Symbian, Microsoft and now Apple. Android is based on Linux and developers can write applications in managed code that use the Google Android SDK, which is developed in Java.

Initially with the Android SDK, there were three ways that a developer could communicate with other services or devices. The first was through text messaging with SMS (or MMS), with HTTP (or the secure variant) or with XMPP, the open messaging protocol. This list has since been struck down to two and a half options, with the XMPP implementation with Android being replaced with a more generic GTalk client library. This means that all real-time communication on Android not using SMS or HTTP must pass through the Gtalk servers.

The initial XMPP implementation was provided via the Smack library, an XMPP implementation developed by Jive Software. Developers started complaining back in January of this year that the implementation had been changed, and that it wasn’t adequate for implementing messaging services on Android. Google has since responded and the official line now is that a more specific library for GTalk is more than adequate, and they also claim that full XMPP isn’t very good for mobile anyway.

Android started as an idealistic open source platform, with a large number of hardware manufacturers lined up supporting it. It is now becoming just a Google platform for mobile, as forcing Gtalk on developers has a chain effect of enforcing Google ID’s on users and all communication to pass through Google servers. Developers are becoming increasingly frustrated by first denying Android SDK updates to some and now by closing up and deciding what can and can’t be done on the platform. Instead of being an open platform, they seem to want to funnel everything through Google services.

As one commentor on the developer thread said, what is next? GHTTP, GSMS etc? Google sucked up credibility by claiming to be the open platform alternative for mobile, but they are burning that credibility up at a fantastic rate with the decisions they are making.