For example, we now think about our pictures as stored on a PC or Mac, uploaded through a USB cable and protected by an online service such as Carbonite or Dropbox. We select which pictures to put in Facebook, Flickr or Picasa Web Albums. In a couple of years we’ll think about the pictures as stored primarily online, directly uploaded there by our WiFi or UWB-enabled camera and protected by a service such as Backupify (yes, we’ll still need backups).

Our computers’ will still have our documents and pictures and music downloads but these won’t be the “master” copies but serve merely as performance optimization and offline access caches for the data, in the same way you can read your email on a plane with no WiFi access using Google Gears for Gmail and Cached Exchange mode for Microsoft Outlook. Amazon’s Kindle is another good example. I read my Kindle books on my laptop, iPad, iPhone, Blackberry and, oh yeah, Kindle. All my content is there, locally cached for fast access without requiring connectivity with the latest read position automatically synchronized without me having to even think about it.

Side note: A few weeks ago Google deprecated Gears with an eye towards HTML5 offering more standard approaches for building these types of application experiences. I would expect significant extensions–going well beyond HTML5–to show up in ChromeOS.

The change will profoundly impact both the hardware and SW industry. On the hardware side, it benefits netbooks and tablet devices as well as smartphones. On the software side, it benefits third-generation SaaS players who are targeting this deployment model as they will have a noticeable performance and user experience advantage over current-generation Web apps.

I’m even more excited about fourth-generation SaaS apps whose user experience will automatically scale up and down with the capabilities of the client-side hardware and bandwidth constraints, similar to the way some gaming engines & services work today.

The shift of master data from PCs and local networks to the cloud creates significant opportunities. Get in touch if you are working in this area–I’d love to hear from you.

1. Convince the market leader to give you something big they don’t see the value in for cheap
2. Innovate, innovate, innovate
3. Make a ton of money
4. Eat the market leader’s lunch
5. Bet on everything big

Guess which is the most important step…

The case of Microsoft

About 28 years ago, IBM made one the biggest strategic mistakes in its history. Being a company primarily focused on hardware, it didn’t understand the value in the PC operating system. It outsourced that work to a, at the time, little-known company called Microsoft. From Wikipedia, the ancient lore of the computer industry:

On August 12, 1981, after negotiations with Digital Research failed, IBM awarded a contract to Microsoft to provide a version of the CP/M operating system, which was set to be used in the upcoming IBM Personal Computer (PC). For this deal, Microsoft purchased a CP/M clone called 86-DOS from Seattle Computer Products, which IBM renamed to PC-DOS. Later, the market saw a flood of IBM PC clones after Columbia Data Products successfully cloned the IBM BIOS, and by aggressively marketing MS-DOS to manufacturers of IBM-PC clones, Microsoft rose from a small player to one of the major software vendors in the home computer industry.

The rest is history. Microsoft took the job seriously, created a pretty darn good product for the time and used the PC clone explosion as a way to increase its dominance in the value chain. IBM lost it leadership. The DOS and then Windows franchises generated massive cashflow which MS plowed into innovation (often in the form of cross-subsidies lasting a decade) to create often dominant product lines such as Office, the server business (don’t forget that Sybase in a no less brilliant move licensed core tech to MS enabling them to create SQL Server) and others. Now, the company bets on pretty much all large markets in the software space. If it’s big, there is no way MS would ignore it. In fact, from conversations with execs there, it seems like much of new market development planning begins with a process as simple as ranking future market sizes. That’s why Microsoft will relentlessly pursue search and why it was probably willing to do an even better deal with Yahoo. (Jason Calacanis has a good post on the Yahoo-Microsoft deal.)

The case of Google

In the case of Google, the strategic mistake was Yahoo’s. The company didn’t value search. It outsourced it to Google. Google took the job very seriously. They innovated a ton behind the scenes while smartly keeping the simplicity of their user experience. Through AdWords and later AdSense + the DoubleClick quisition, Google gained significant control of the advertising value chain. They passed by Yahoo and never looked back. The search franchise is so profitable that it gives them freedom to bet on everything from browsers to OSs to mobile, something pretty much no other company save for Microsoft can afford. Google’s version of the cross-subsidy is the perenial beta that just keeps getting better. Very, very hard to compete with that.

The moral of the story

The moral of the story is that the two most important software companies in recent history were enabled by massive strategic errors on behalf of the market leaders in their respective sectors. It was the market leaders who put the companies on the map, gave them distribution and scale and enough revenue to fuel innovation. IBM and Yahoo nursed the companies that nearly killed them. One could argue that without the explicit support of the market leader, a startup wouldn’t have had the chance to grow at a rate that allows it to have Microsoft’s and Google’s escape velocity.

Kudos to IBM for innovating around its business model and evolving into a company which is still a huge industry player. I’m not sure Yahoo will have that chance.

The interesting question is whether this will happen again… Do you think a market leader will make that type of mistake again? Can you point to where this is happening now or likely to happen in the future?

The AIR / Flex 3.0 launch is a big deal for three reasons:

• The AJAX ecosystem is heavily fragmented
• Flex 3 is a big step forward
• AIR is a disruptive innovation enabler

There are just too many pure AJAX frameworks out there. This may be great for some abstract notion of innovation but is pretty bad for market adoption. It causes confusion and friction. Developers hate that. They want to feel like they are developing skills for the next cool technology which will be around for a decade. The tooling is also pretty immature, although my friend Paul Colton @ Aptana is doing some pretty cool work. I have also seen a number of new startups trying to make things better in this space but for now the situation is not great. After initially missing the AJAX boat in a big way, Adobe has recently gained traction in repositioning from Flash vs. AJAX to Flash vs. DHTML, both of which use AJAX. As proof, I’ve seen more startups mixing Flash/Flex with DHTML/AJAX.

There seems to be a rule that really big software companies get things right by version 3.0. That has been the case for years with Microsoft and now Adobe may be large-enough to join the list. Flex 1.0 shipped late and left a lot to be desired, which made Flex 1.5 more or less a requirement. Flex 2.0 was a big step forward but there was much more left to do to build a great RIA platform. Flex 3 steps up the innovation pace on a number of fronts. Better developer tools. Improved frameworks. Smaller application size. Better data handling throughout. And the Flex SDK as well as the BlazeDS remoting and messaging services are now open-source. Everything is in place to accelerate Flex adoption. There is a massive shortage of Flex developers in the market, always a good sign for a platform. The main challenge for Adobe remains monetizing the platform. I’m not 100% on board with the Flex Builder tool as the main revenue driver although the team recently added some great people from the ColdFusion team who really understand Web developers. For examples on what Flex can do, check out the app showcase. Warning: you may spend more time there than you previously  imagined.

AIR will become the biggest enabler for media-heavy Web applications. To get an idea about what I mean, check out the Allurent Desktop Connection video. This only scratches the surface of what one can do with a cross-platform, always on, richly interactive application with Net access and plenty of local storage. Adobe isn’t the only company who gets this–Google has plans based on Google Gears and Facebook has also moved in this direction with the Parakey acquisition–but Adobe has so far executed the best. Not surprisingly, Microsoft remains paralyzed by the Web <-> desktop strategy tension.

AIR is truly disruptive–in Clayton’s exact sense of the term–for two reasons. First, it given Web apps access to the desktop. Second, it does so while subverting the native operating system installers (using the Flash player executable as a conduit) which gives Adobe and its ecosystem substantial freedoms. One caveat–security. Adobe has done a pretty good job with the Flex 3 security model but it hasn’t been extensively tested in the wild. Also, the anti-virus and anti-spyware vendors have often been trigger happy in the past to blacklist totally legitimate software. I’m not expecting them to blacklist the AIR runtime but who knows how they’ll treat some of the apps… I hope the Adobe platform marketing team and evangelists have spent enough time with the security vendors. This is something I advised them to do at the Engage event last year based on some bad experiences from our portfolio.

With this launch, Adobe cements its platform leadership in Web 2.0 and makes life easier and more interesting for millions of developers and designers. A shout out to all my friends at Adobe who worked hard to make it happen and to everyone who put skin in the game during the beta process.

The weather this year was both great and marginal. Deer Valley got about  55-60” of snow in four days. The powder made the skiers happy. Those of us who don’t do the downhill thing went snowmobiling. We had a good time although snowfall killed the visibility as you can see on the pic. On the flip side, getting to dinners with the road closings was a challenge and today a few attendees narrowly escaped an avalanche at Deer Valley Resort.

We covered many important topics: from how to build organizations in digital media startups and scale infrastructure, to new content & advertising models, to monetization of UGC and other forms of social media, etc. Many of the comments were insightful due to the diversity of backgrounds and perspectives around the table. I have a few good follow-on topics for blogging that I hope to organize and post in the next few weeks.

On Friday the group was abuzz following the news of the Microsoft offer to buy Yahoo. Several of the attendees had stayed up overnight and were prepping for board calls and press interviews. Microsoft does best when someone is breathing down its neck. I hope the same is true for Google.

As an ex-platform guy who’s recently spent a lot of time thinking about social infrastructure, I’m was particularly interested in the opening panel on emerging online platforms, moderated by David who kept the conversation moving swiftly. The panelists were:

• Marc Benioff, CEO, Salesforce.com
• Marissa Mayer, VP Search & User Experience, Google
• Mark Zuckerberg, CEO, Facebook
• Philip Rosedale, CEO, Linden Labs (maker of Second Life)

The discussion centered on what it means to be a platform company on the Net and to get people to use your platform.

Philip and Mark acknowledged that they are in the very early phases of truly understanding what it means to be a platform yet there are some underlying principles for success.

Philip stressed that it is foolish not to allow your users to do what they want. Sometimes, that requires making some seemingly uncomfortable decisions. Traditionally, platform companies have tried to protect and leverage the core platform IP. However, if a platform enables network effects (which make it harder for users to leave), the platform owner can benefit by opening up and applying only a few key controls. Mark pointed out that there is a huge issue of trust between users, the platform provider and third party developers. That’s one key area of controls that the platform vendor should maintain control over for the benefit of all.

There are at least two other areas that IMO the platform provider should establish clear rules, if not controls, over.

The first one is around the question of who owns the data. Esther Dyson asked the panel a question along these lines and nobody provided a satisfactory answer. Mark’s response was “users have all the controls” but that doesn’t answer the ownership question. I’ve written about this before as one of the key issues underlining the emergence of social infrastructure.

There has been much debate recently on the topic of user data ownership. The information in question is the user-generated data and the attention data generated while using a site. There are three issues to consider. The first one is legal. Users implicitly assume that the profiles and the information they create on community and social networking sites is theirs. Most terms of service and privacy agreements are either mute on the topic or outright state that the operator can repurpose the data for their own use. Attention data is often not called out and its treatment therefore is unclear. The second issue is privacy. Facebook recently faced a huge backlash from their users when they made information about user activities on the site public. The specific outcry was less about the features Facebook had implemented and more about the fact that the information in question was gathered while the users thought it would be kept private. The company had to backpedal and put new privacy features in place.

The second one has to do with the notion of who is the law. Who provides the framework of what’s good vs. bad, what’s legal vs. not. Is there a notion of enforcement? By whom? If there are spillover effects into the real world (financial fraud, privacy violations, etc.) how is the real/virtual linkage managed. Release 2.0 had a good piece on this topic in its last issue.

Marc’s take is that the Net has become the OS and the killer apps (Second Life, Facebook, Salesforce.com, etc.) are becoming platforms. As an aside, none of these have any dependence on the “old” platform–the Microsoft OS. He further claimed that killer app companies that do not make the transition to platforms become irrelevant. His examples were Seibel vs. Salesforce.com, Oracle vs. Cullinet, etc.

One interesting question that comes out of this is whether one can build a successful platform company without ever building a killer app. Would Windows have been as successful without Office/Outlook? Unlikely. Could Facebook have launched as just a set of Web services? I bet not. I can’t think of many examples where a platform company has been successful over a long period of time without some ability to create user lock in + lead by the example of its killer apps. My personal experience at Allaire, which was a pure platform company, agrees with these observations.

In addition, everyone agreed that platforms only become successful if they give rise to successful ecosystems. One of the metrics Philip is watching is how many businesses in Second Life are cashflow positive. Facebook now has thousands of apps on its platform and some of them have millions of users. Marissa’s work at Google has enabled a huge monetization of content and traffic on the Net by many thousands of businesses. Beyond that, Google has been very successful with Maps, iGoogle and gadgets (10,000+ developers). I’ve written before about the Ecosystem Test that a platform must pass in order to be successful.

The ecosystem test asks whether a platform can enable a large group of average, poorly funded players with little to no domain experience deliver compelling solutions. It’s based on the observation that no platform has become hugely successful without a corresponding ecosystem of vendors building significant businesses on top of the platform. Typically, the combined revenues of the ecosystem are a multiple of the revenues of the platform.

The panel was followed by a short speech & interview with Catherine Cook, co-founder of myYearbook. She sees myYearbook, Facebook and others establishing the infrastructure for a new wave of media companies that will give old media a big challenge. One example she gave was myMag, a 100% UGC teen magazine they’ve launched. Teens contribute the content but professional editors package it.

Let’s look at some data.

• Symantec bought
  • Veritas (July 2005), which bought KVS, an email archiving and discovery company. (Aug 2004) KVS’s technology has been packaged into managed service offerings since 2004 by companies such as AT&T.
  • IMLogic, an IM security/archiving company (Jan 2006)
• Iron Mountain, once a boxes-on-trucks-and-in-caves company, bought online backup player Connected (Oct 2004), put its leadership team in charge of Iron Mountain Digital and set out to create the most scalable outsourced digital content discovery and archiving platform.
• Autonomy, one of the key players in enterprise search, bought discovery and archiving vendor Zantaz last week for $375M, which in turn had bought hosted e-discovery vendor Steelpoint. (Aug 2004)
• Yesterday Google announced it will acquire Postini for $625M, a hosted content discovery, security and archiving player.
• Even Tumbleweed wants to buy a hosted email archiving vendor.

Some trends emerge:

• Outsourcing of content archiving, which often means
• Outsourcing of e-discovery, which is why we see
• Search players buying hosted content archiving companies

To move content off-site, hosted backup/archiving providers need to tap into all key enterprise information flows (network, email, storage, desktop, etc.). This allows them to offer value-added services such as content security (DRM, IP theft protection, etc.) as well as implement a number of compliance-related policies. In addition, once content leaves the enterprise, there are many possibilities for additional services around e-discovery, collaboration, data mining, etc. That’s why everybody wants the enterprises’ data.

The biggest challenge to implementing this well is tapping corporate desktops–where most of compliance-related information (emails, docs & spreadsheets, IMs) ultimately originates. (Integrating with servers is easy + there are far fewer of them.) For a solution to truly work and be deployable to employees without requiring costly training and change in work habits it will have to be seamless–practically invisible and non-disruptive to the unstructured workflows of knowledge workers. There are two easy ways to do this. You can build these capabilities inside the office productivity software. Microsoft and the folks at Open Office have the lead there. You can host the productivity apps, which is Google’s long-term strategy for the enterprise. It is a very powerful idea, indeed. Right now, I know of only one solution on the market that can tap into MS Office & Sharepoint environments without disrupting employees. It comes from Meridio, a Belfast-based company I’m on the board of.

Some interesting questions:

• Does IBM have this covered through it’s outsourcing offerings and a myriad of ECM technologies?
• When will Microsoft get into the hosted content services business in a big way? Nearly 1/2 of KVS’s business was on the MS platform…
• How will EMC respond? They have pushed to increase their service offerings (bought Internosis in 2006) and have been integrating search deeper into their storage and ECM product lines. Perhaps they’ll buy Autonomy?
• Will FAST stay focused on search alone?

The pitch of the acquisition announcement is that, by adding the Postini services to their portfolio of offerings, Google’s enterprise group will be able to address many of the concerns of larger enterprises not willing to move to Google’s hosted office productivity applications. I don’t buy this as the short-term reason for the acquisition. In the short run, larger enterprises won’t switch to the enterprise edition of Google Apps. I’ve been using Gmail, Google Calendar and Docs & Spreadsheets for a long time now. They don’t have the usability, offline capability and interop features to battle MS Office or Open Office now. I’m sure Google Apps will get there eventually with the help of time and Google Gears but not in the next 12-18mos.

The longer timeframe of extracting value from the Postini acquisition is one reason why I consider the acquisition strategic. The other is that through the acquisition Google is getting their hand on the content of 35,000 businesses and 10M users. That’s great for search and an important tool to help make Google Enterprise Search the winning product in this market. They have a long way to go, but the Postini move is a step in the right direction.

• Microsoft buys aQuantive for $6B (2x price premium suggests some deal competition)
• Google buys DoubleClick for $3.1B (beats Microsoft)
• WPP buys 24/7 Media for $649M (Microsoft was also looking at it)
• Yahoo buys Right Media for $680M
• Just for kicks, in the mobile advertising space, Microsoft buys Screen Tonic and AOL buys Third Screen Media for somewhere between $105M and $175M (apparently, Google was also interested in buying the company)
• To top it all, Microsoft fails to buy Yahoo.

It feels like a game of Monopoly where the three players with a lot of cash (MS, Google & Yahoo) are buying up all the real estate. Across both the Web and mobile, control of advertising is consolidating. Search has long consolidated. If search is the traffic driver and advertising is the money maker, then it is difficult to escape the conclusion that the Net is increasingly falling under the control of the majors.

Only Google and MS have guaranteed staying power for the next few years. Since the fate of AOL and Yahoo is yet to be determined, there is a possible future with an overwhelming, antagonistic duopoly controlling the majority of search traffic and ad delivery.

This will be an interesting world for both startups and big publishers, one requiring significantly adapted strategies for success. It’s not all bad news. Sometimes restricted but efficient choices enable better outcomes.

Using Occam’s Razor as a guide, this naïve explanation may just be why it made sense for Google to outbid Microsoft for DoubleClick. I discount the economics that DoubleClick can bring to Google for two reasons: (a) some of DoubleClick’s current customers/partners will look for alternatives and (b) Google is still learning how to handle large acquisitions. Also, from an economic standpoint, Google would not have been threatened in the short or medium run by DoubleClick going to Yahoo! or Microsoft. However, the opportunity cost of losing DoubleClick from a perception standpoint was huge. It would have exposed a crack in the armor.