Turning the Tide

The other day I was trying out IE 7. It has a nifty feature that downloads all the elements of a web page and stores them together in one file for future viewing. We do a lot of work with SharePoint at my company and I was curious how many bytes a generic, freshly created Document Workspace takes. It turned out to be 715K or so, which is surprisingly close to a megabyte. Thinking maybe this was MS/SharePoint specific, I tried a new Google Docs page (472K), a new Google Spreadsheet page (418K) and the front page of Yahoo Finance (429K). Somewhere along the lines, as HTML pages have entered the mainstream as real workhorses, the nimble HTML burros of the early web have morphed into plodding, three-tiered-architecture clydesdales. It doesn't take too many clicks and page refreshes before you've downloaded more bytes than a comparable rich application. Many, many more bytes. And that's just one page.

Could history repeat itself? What will be the tipping point for a few rich application machines to replace scores of web page clydesdales? One possibility is an increase in the cost of all those bytes, which is why Google (and most other web application providers) really would like to see a Net Neutrality amendment or bill [update: we edited the previous sentence after publication, based on comment 14 below].

More likely is something like Ektasis, a startup company that recently introduced itself to the world. Building a new platform, as noted earlier, is difficult enough. Handicapping it with a startup's cash constraints is insanity. But it's folks like these that make the world a better place. And thankfully, they do have a very impressive solution to the install/uninstall problem facing rich applications today. It's a solution that Microsoft would do well to copy and Google to study: fully functional client software that can install and run with a click, and an automatic code versioning/updating system. This means instead of downloading entire applications, pieces of applications - individual classes or object files - can be retrieved and fine-grained, surgical updates can be performed. What they've done is taken the best practices of browsers and web page assembly - and applied it to desktop apps. Ektasis has a great strategy for merging the desktop and web environments. Perhaps they will be able to catch the right wave.

Any virtual machine can also use this piecemeal download and assembly strategy. In theory, an operating system like Windows could do it too. But in practice, because of its own success, there are too many millions of windows applications to support - some written by software publishers, most written in offices and cubicles in every part of the world. Therefore the better solution is to lay new groundwork for a better virtual machine, something that includes a smart loader/linker like the Ektasis framework. This will negate the installation/uninstallation advantage web applications enjoy today. Furthermore, it will give developers more choices in how they want their apps to work.

More choices? The other nice thing about virtual machines is that they are portable, just like a browser. But it's not for running .NET apps on Macs, though that may happen. Rather, portability is gaining in importance for mobile devices. Any company with an eye for growth has noticed cell phone sales far surpassing PC sales. Add millions of cell phones, Blackberries, SideKicks to the environment and it's clear the playing field of tomorrow is a whole lot bigger than the playing field of today - if you're able to span devices.

Google's Dilemma: The Next Big Thing?

As enormous as Google seems, its position is far from unassailable and their options are surprisingly limited. There is nothing preventing Microsoft from duplicating Google's online strategy, other than prime mover momentum. Even with copious advertising profits and their astounding growth rate, Google makes only a little more than 10% of what Microsoft makes. Which means MSFT can easily outspend GOOG. Enough to slow Google's current momentum? Probably not. Enough to pounce on Google when advertising profits suffer a downtown? Most definitely yes.

In fact, Google 2006 reminds me a bit of AOL in 1996: long on one cash cow, short on any others. In 1996 AOL was the growing giant. By 2001 the giant's food source had changed. The cause of AOL's problems was over-reliance on their growth engine. Once it started sputtering, they had no other profitable property to rev up. As a result, even though they had well known properties (AIM, 'You've got Mail', etc), they could not monetize them and thus were dead in the water. 

For all the apps Google puts out, very few can be considered 'sticky', and even fewer can be monetized outside of advertising - it's too easy to go to Yahoo or MSN for the same free service.

Indeed, Microsoft is already encroaching at each potential feeding ground. Google offers Earth, Microsoft offers Virtual Earth. Google offers AdWords, Microsoft offers AdCenter. Google offers documents and spreadsheets, Microsoft offers Office Live. Google invests heavily in Firefox, Microsoft shakes the dust off Internet Explorer. Google has built its momentum by brilliantly exploiting the web. While Microsoft just seeks to match them, they have a fighting chance. But what happens when Microsoft takes the battle to the next level and introduces a smarter, portable virtual machine that unifies the development experience for PCs and mobile devices? Google answers with... what?

Perhaps they can invest more in Firefox. Perhaps Eric Schmidt fondly remembers his Sun days and resuscitates JavaOS (he did just resuscitate 'The Network is the Computer,' after all). Unfortunately for Google, Microsoft is more than a match for them on both counts. One possibility is buying an undervalued property aligned with its business, but currently lacking frothy sizzle. A company like EMC (an information infrastructure company). It aligns well with Google's information driven goals, has lots of assets and even more relationships. But even that may be window dressing for EMC's crown jewel, which might be worth the acquisition alone: VMWare.

Acquisition options for Google

VMWare has actually, unbelievably, beaten Microsoft on its own turf - building a better virtual machine (the V and M in VMWare). I'm sure they could create virtual machines that would run .NET and .NET apps. While Google would never be able to root out Windows on PC machines, they could capture the bulk of billions and billions of cell phones, mobile and personal devices that will deluge us over the next several years.

Such a staid purchase, however, might wash a bit too much glitter off Google. For something a little more snazzy, and just a bit more affordable, they could make an even bigger splash and buy Adobe. It's all funny money for Google at this point and buying Adobe would give them fabulous software assets in PDF, Flash and Photoshop - three pivotal areas of the web, and markets Microsoft has been unable to capture. Furthermore, Adobe's ambitious Apollo project could become a crown jewel, too. Finally, Adobe is just up the creek from Mountain View. (Full Disclosure: I'm a former employee and own shares of Adobe)

But Google will have to act soon. Microsoft has already taken away one potential option for Google - Novell's .NET Mono project. Microsoft must be thinking about the mobile deluge too. They've already met Google head on at every online PC location, and there's every reason to believe the same thing will happen at every mobile hotspot as well.

If Google wants to avoid being the next AOL and instead make a real grab for software supremacy, it needs to expand and diversify its revenue source. Fast. Firefox is not enough; their own OS is not viable. With their stock price Google has some options. Perhaps a solid information infrastructure provider like EMC, or a 24 carat technology treasure like Adobe. Either way, Google needs continued brilliance at managing their search armada, stickier internet properties they can monetize and an extra boost to propel their ascent past Windows.

And they need Microsoft to make a mistake.

Windows is Dead. Long Live Windows.

Have you ever read Slashdot and noticed that every article about Microsoft is accompanied by a picture of Bill Gates made up as a Borg? While it may appear in bad taste, it's actually a very flattering compliment. For most of their history, Microsoft has been able to adapt like very few companies before it - and it speaks to the very core of their success. Though the software environment may be warming in Redmond - the water lapping at their ankles and the investors clamoring for greater returns from the hunt - Microsoft has usually been able to make hard, but correct, choices. And when they don't - then scariest of all, they learn from their mistakes. The result is that Microsoft has a solid position at the top of the software food chain.

Learning from their mistake with Linux is what led to the recent deal with Novell. Microsoft's initial mistake? Not taking Linux seriously. Now it's a serious headache for them. Though it may seem a stretch today, another headache had been looming on Microsoft's horizon - especially with the growing importance of .NET. That was Novell's Mono project. Mono allows .NET code to run on any machine that can run Mono, similar to HTML code running on any machine a browser runs on. But Novell needed the money; Microsoft needed to remove a threat. The solution: pay chump change (for Microsoft) to Novell for access to Mono. Lesson learned and problem solved.

But just as Google 2006 reminds me of AOL 1996, Microsoft 2007 will soon have a major decision to make similar to Microsoft 1997. Back in 1997, strange as it may seem, Java was the rage and elements within Microsoft were comparing and contrasting the virtues of virtual machines with the ungainliness of Windows. A management shakeup resulted and Microsoft, correctly, continued to orient itself around their OS and fortified any cracks in the foundation. Fast-forward to 2007 and Microsoft will again be comparing and contrasting virtual machines with an even more unwieldy Windows.

This time however, Windows is acting like a dam blocking a surging ocean of innovation. Further fortifications are useless - you can't stop a sea change in technology. Of course this time Microsoft happens to own both options on the table: a nimble virtual machine that can run on as many devices as needed, or an unbowed warhorse ready to fight the last battle. Once again it's the horse that must go. Not Windows the brand, but Windows inextricably tied to the PC platform.

Except in 2007, it won't be Bill Gates delivering the memo. The mantle of technical leadership, and compelling memo writer, has fallen on Ray Ozzie's shoulders. His greatest test? Slowly defocusing Windows bound to a PC and refocusing on a portable, virtual machine 'Windows' fueled by .NET and online services. My company worked with Ray's former company, Groove, for many years and while I haven't had a chance to sit down with him, I know he's very capable technically. But it's the selling part - the clarity of vision, the sureness of direction, the respect of employees, the sheer force of personality - that is extremely difficult for 99.9% of the world. In that respect, Bill is an exceedingly tough act to follow.

How important is it to have a leader when coming to a fork in the stream? History tells us that most empires start crumbling from within before the outward edifices are breeched. Microsoft head count has nearly quadrupled since 1997, as has revenue. There's a good chance the number of internal fiefdoms has quadrupled - as well as competing interests. It was definitely surprising how little the web was mentioned in the recent Vista/Office 2007 press event. Perhaps Google's best bet is whispering sweet nothings in every willing ear.

When the Storm Clears

But for all the managers, marketers, salesmen and saleswomen hired, Microsoft is still, at its core, a technical company. In order for the outside world to know if the technocrats at Microsoft have bought in to Ray's vision, the key development to watch will be 'old' Windows being put out to pasture. Again, Windows the brand lives on; Windows the tightly integrated OS for x86 computers does not. It's not a unique situation, it's just that the stakes have never been higher. DOS was a tremendous cash cow before Microsoft replaced it with the greatest cash cow in history: Windows. There's every reason to believe the next cash cow will be even greater. If Microsoft engineers can convince everyone it's time for Windows, following the revolutionary trend, to disappear and be re-imagined as a virtual machine intricately tied to the web, then Google will be up a creek no matter what they do.

Remember, you always want to follow the data. Microsoft and Google are struggling to own it on the estimated 234 million PCs shipping this year. But it will be the first company that can extend their reach to the 245 million mobile devices shipped last quarter that will be the winner. While Google has the richer feeding grounds as Microsoft struggles with the current Windows/Desktop status quo, it's actually Microsoft with the canoe, the paddles and the most rods and reels. If they get everyone on board, then Microsoft should continue ruling the land and the seas.

Until the next unintended consequence.