The Keene View on Web 2.0, Ajax and SaaS

McKinsey & Company published a report predicting the market size for Software as a Service (SaaS) will exceed $37B market over the next 5 years. In particular, the report described the need for Independent Software Vendors to SaaS-enable their products using special-purpose SaaS development tools. Matt Asay also wrote recently that the growth of the top 60 software companies is driven by SaaS.

McKinsey claims that traditional J2EE and .NET platforms are poorly suited to building SaaS applications. According to McKinsey, this opens up a $3B market for Platform as a Service (PaaS) products from new entrants like WaveMaker, Coghead and SalesForce. From the article:

Although SaaS development platforms like SalesForce and Coghead have gotten a lot of attention, this market has so far been remarkably closed and proprietary. The Platform as a Service leader, SalesForce, has both a draconian hosting policy (host your apps and data anywhere, as long as it’s with us!) but also a proprietary language (who needs Java when you’ve got Apex!?).

Moving forward, the same trends driving open source adoption everywhere else in the industry will ultimately drive SaaS adoption of open source, particularly by ISVs whose business plan does not include a low multiple sale to their proprietary hosting provider. Future SaaS platforms will converge with traditional tools, offering on-demand development based on traditional programming languages with built-in tools for mash-up based development for basic users.

Development Problems for SaaS

SaaS is highly disruptive for existing hardware and software providers. SaaS platforms are different from traditional computing platforms like J2EE and .NET in three ways:

• SaaS platforms contain new core components, such as web services APIs to integrate to other applications and usage-based billing capabilities. This disrupts existing platform providers like BEA and Microsoft.


• SaaS platforms are designed for multi-tenancy, including global and tenant-specific data schemas, multi-layer administration and virtualization for scalability. This disrupts traditional ISVs like Oracle and SAP.


• SaaS platform are delivered on-demand, not on premises. This threatens the business of traditional hardware providers like IBM and HP.


• SaaS products need on-demand customization tools. As SalesForce has demonstrated, a complete SaaS application needs its own customization tools if it is to compete with enterprise solutions like Siebel and SAP.


• SaaS products need on-demand integration capabilities. This includes ability to integrate with on-premises data (a notorious weakness of pure-cloud solutions like Force.com) as well as with on-premise and on-demand web services.

SaaS Architecture Requirements

McKinsey identified three elements of a SaaS architectures:

1. Development environment: an on-demand development platform for creating SaaS applications. This platform should be able to ship along with the application itself to allow customers to customize their application.


2. Run-time environment: an on-demand infrastructufre to deliver applications. This can be a proprietary hosting environment like SalesForce, or an open hosting environment like Amazon EC2. Ideally, the customer should be able to deploy applications on-demand or on premises depending on their security, data integration and other requirements.


3. Ecosystem for adding new capabilities to applications (e.g., SalesForce AppExchange). This ecosystem should also be able to access enterprise data and services located inside the enterprise firewall.

SaaS Is Make or Break for ISVs

According to McKinsey, SaaS has greatest impact on ISVs, delivering a 50-70% improvement in the level of features that can be delivered for a given investment in development and infrastructure.

For ISVs, SaaS platforms offer low upfront cost, rapid time to market (productive tools + pre-built components like billing) and high quality service delivery. In short, existing ISVs have a limited window to migrate their offerings to the SaaS platform or risk being obliterated by newcomers who get there first.

The lesson of SalesForce versus Siebel Systems is clear: existing ISVs should migrate their presentation layer to SaaS quickly while preserving their existing back end servers. Preserving existing back end logic requires a SaaS platform that supports traditional languages like Java.

Which Platform Will Win the ISV Business?

A battlefield is emerging between established mega-vendors and pure play SaaS vendors. The following factors will separate the winners from the losers in this market:

• Build a robust offering: cutting edge technology, reliable, high quality.
• Enable extensive customization: provide additional components that address SaaS-specific needs (e.g., authorization, billing, monitoring & management).
• Monetize effectively: McKinsey identifies this as the most important success factor. The winning platform vendor will be the one which most effectively creates economic value for its ecosystems!
• Drive ecosystem growth: enable partners to make money within the platform vendor’s community through collaboration, sharing of tools and best practices.

Although many of the early SaaS platforms are based on proprietary languages and tools, Gartner predicts that 90% of SaaS software will be based on open source within 2 years.

Evaluating SaaS Platforms For ISVs

Here are important criteria for ISVs to consider in evaluating SaaS platforms (sometimes called Platform as a Service, or PaaS):

• Open hosting: can I move applications I build to another SaaS hosting providers? Many SaaS platforms lock the ISV into a proprietary hosting provider (e.g., SalesForce). ZDNet says that ISVs need to offer their SaaS software both on demand and on premises.
  
• Full platform: does the SaaS platform offer a complete development solution with presentation layer, business logic, security, database and web services? Some SaaS platforms only offer part of the development stack (e.g., DabbleDB, Tibco GI)
  
• Standard language: does the SaaS platform support development using a standard language such as Java? Many SaaS platforms are based on proprietary languages (e.g., Apex, the proprietary language for SalesForce).

Table: A Comparison of PaaS Vendors

* Proprietary language

Peter Laird also has a good SaaS platform review and Phil Wainwright’s has a good comparison of PaaS providers.

SaaS Platform Product Review - WaveMaker

WaveMaker is an open source, visual development platform for building Web 2.0 applications. The WaveMaker studio can be installed on a developer workstsation or delivered on-demand. WaveMaker creates standard Java applications based on Spring, Hibernate and Dojo that can be deployed in a SaaS or on premise architecture.

For ISVs, WaveMaker offers several compelling benefits:

1. WaveMaker's visual studio provides a faster and more natural way to build rich internet applications than traditional hand-coding using Java and struts
   
2. WaveMaker is completely open, making it portable across hosting providers and even enabling applications to be deployed on premise
3. WaveMaker includes a complete development platform based on open source standards such as Spring, Hibernate and Dojo
4. WaveMaker is based on the Java language, making it an ideal choice for ISVs who already develop in Java and don't want to migrate their existing server code.
   

WaveMaker can be downloaded here.

Summary - What ISVs Need From SaaS

Every ten years there is a dramatic shift in the development tools world: in the 80’s to client/server, in the ‘90s to three tier and now in the 00’s to SaaS. In each of these shifts, the dominant development tools providers have been supplanted by a new generation. This time around, the seismic shift is being driven by the on-demand architecture and the ISVs have the most urgent need to rebuild their solutions to remain competitive.

Over the next five years, we will see the 500 pound gorillas of the development world like Microsoft’s ASP.NET and Sun’s J2EE unseated. In their place will be new software platforms based on traditional languages that are specially designed to enable development of SaaS applications.

Labels: paas, platform as a service, saas, salesforce, software as a service, WaveMaker

I have been on Facebook for 3 months and although my next comment will instantly brand me as tragically unhip, I find it a complete waste of time. I keep on expecting some sort of mystical Web 2.0 insight if I just stick with it for another week - instead, I just get more confirmation that Facebook is more of a step backwards than a step forwards.

This last week brought another blogger-driven Facebook tempest in a teapot when Robert Scoble tried to synchronize his Facebook contacts with his Plaxo contacts. Facebook shut his account down, drawing howls of protest from Kara Swisher at the WSJ and a typically thoughtful rebuttal from Nick Carr. By the way, Facebook already supports a one-way import of gmail contacts into Facebook.

There are two interesting points from an Enterprise Web 2.0 perspective here:

1. There's no lock-in like SaaS lock-in. Software as a service offers a spectacular downside in the case that the service provider doesn't like the way you are trying to extend their service. At its most extreme, closed SaaS is the Roach Motel of enterprise software - your data and logic may check in, but they're never coming back out, as I wrote here.
   


2. Scrape-ability will be an increasingly important battleground: I sit on the board of Kapow, a company that has powerful tools for gathering data from public web sites. If those web sites block access by bots (this is what happened in the Scoble kerfuffle)

In short, much of what is presented as Web 2.0 magic is really just lipstick on a tired old Web 1.0 pig. Here are three examples:

1. Microsoft Silverlight: everything bloated and retro about Windows brought to Internet Explorer. Silverlight, despite its technical innovations, is yet another attempt to assimilate the recalcitrent web beast into the Microsoft borg.
   


2. Adobe flex: everything proprietary and designer-wonkish about Flash brought to any browser. Given that enterprise developers are more interested in formatting data than getting their company logo to spin and then fade, making the design-heavy Flash language an enterprise standard is a lost cause.
   


3. Facebook: if you liked AOL, you're gonna love Facebook! Facebook represents a cautionary tale for enterprises looking at SaaS solutions like SalesForce and Netsuite - make sure you have an ironclad way to get your data and logic back out!
   

ps to be fair, I do enjoy one application on Facebook, ilike, because it tells me when artists like Ryan Adams and Emmylou Harris are playing in San Francisco.

Labels: facebook, Flex, kapow, karaswisher, nickcarr, robertscoble, salesforce, Silverlight