Last month, I attended a Software as a Service two-day conference hosted by Madrona Venture Group here in Seattle. There were some very good round-table discussions with the CEOs of NetSuite, Concur, and 37Signals.
It really hit home for me how we're in the midst of the creation of the Web Operating and Marketing System. And this OS is "moving up the stack" every day, progressing from basic services to high-level application-layer services relatively quickly.
The "stack", in computer terms, is usually represented as a series of blocks, one on top of the other, signifying the layers of functionality. The stack was largely confined to hardware and single-user input/output tasks for 25 years of the PC's life. It had a "silicon ceiling" if you will -- the model was that the O/S would take care of talking to the devices and the hardware, and leave the application-level coding to the application developers, who would very rarely try to interconnect those applications.
But now, all of a sudden, the silicon ceiling is shattered, and thousands of startups and large companies alike are racing to engineer the next application and marketing layers of the platform. This work is largely going on in parallel, not in series. It's being done not by a few monolithic companies, but thousands around the globe.
1970's-90's: The desktop PC stack
At the lowest level is the hardware. Electrical signals can be sent into the CPU and other electrical signals come out. The magic that makes a computer work is at the BIOS (basic input/output system) and the operating system. The desktop PC operating system stack tackles things like writing to the display, disk input/output, temporary RAM storage input/output, connecting to peripherals like printers, mice, network adapters etc. As it got more advanced, the screen writing and networking became far more sophisticated, but essentially, it was still dealing with basic I/O and memory management tasks.
OS's are natural monopolies. Why? Consumers want powerful, inexpensive applications, and developers seek the largest market for their development efforts, since in most cases, they need to "bet" on an O/S to get the most power out of it. As a result, economically, OS vendors receive great lock-in and network effects.
The Web Operating System will have similiar economics -- those who do the best job serving as layers in the cloud can extract economic rents, though perhaps not if they become too commoditized through standards. I believe that the application-layer services in the stack have a real opportunity to provide tremendous value to customers, and also build lasting businesses.
Today's Web Operating System -- The "Cloud" plus Marketing
There's been a lot of talk in tech circles about "the cloud" -- this is analagous to data-management, communication and application-layer operating systems available on-demand.
In the past 5 years, the Web has moved from simply providing web browsing and file-serving (simply fetching data from web servers) -- to the next logical layers of the stack -- the application layer. What lives there? Things like:
- Social operating system services: Single log-on, central address book storage, calendering, list management, instant messaging, etc.
- Search
- Image and video hosting, communication and manipulation
- Financial services -- accounting services, payroll services, ERP services, etc.
- Secure, fast database storage and retreival
- Monetization (advertisement, subscriptions, etc.)
- etc.
Importantly, and new for most developers, is the fact that Marketing Networks are now becoming a significant decision early-on in the life of a company. Should you write your business on Amazon.com Web Services? Or Google's App Engine? Or Salesforce's Force.com? Or NetSuite's platform? Or Facebook's platform? Or go it alone? What are the benefits and risks of each?
For developers of software-as-a-service (SAAS) applications, the decisions start to look a lot like the ones console videogame developers have been making for years. Do I bet on the XBox platform or Nintendo's Wii, or the nextgen Playstation? If I do, what's in it for me? I might get some really great marketing and network-effects, but only if I'm a premier app. To gain the most marketing and "bundled promotion", I'll need to write my game to be a flagship for Microsoft -- I'll need to show off particular differentiating features of that XBox network. So too for Google's App Engine, the iPhone, or whatever platform or device you're writing for...
The Cloud is starting to come with marketing services and broad-reach networks.
Implications for Microsoft
Strategically, Microsoft needs to internalize that Search is just one of the functions of the Web Operating System. The long-lost "Hailstorm" project actually had it right back in 2001. Perhaps Microsoft isn't in the brand position to build and power it, but it's coming.
Microsoft knows how to build platforms, and platforms are being developed right now. Why isn't Microsoft the leader in free API's for search? Why isn't it possible to look up an API on Microsoft's site for just about anything you'd like -- to manage subscription payments easily, to send money via email, to get a full Plaxo-like self-synchronizing address book for anyone in the world, etc. Why the hell doesn't Expedia (note, not Microsoft-owned anymore) or any other travel provider let me search and buy airline tickets through an API?
It's not just about winning today's search war. Sure, it's extremely vital, but in the end, the Web Operating System dominant players will be the ones with the most complete suite of easily "mashable" services and the highest lock-in of developers and users. This, in the long-run, only holds together by delivering the highest value. And ultimately, I believe application developers will force more and more API's and openness of that system. Microsoft and their value-network benefitted greatly from the PC operating system revolution, and was significantly better at courting developers than other OS vendors, but ultimately is being forced to be even more open.
The recent moves by Yahoo to open up their search (for both input and output) are an interesting foreshadowing of the next war ahead in the Web Application Operating and Marketing System.
Comments