The SaaS Steamroller

Last month’s article, the first in this series, discussed the evolution of software as a service (SaaS) from the application service provider (ASP) model to today’s SaaS poster child,

It discussed the forces making some companies re-think their approach to delivering IT services to the business, and move toward a service-provider, rather than an in-house, delivery model. Finally, it discussed, some of the advances making SaaS a viable business model where ASP was not.

This article details the reasons why SaaS is different, and why it will likely begin to impact enterprise software vendors over the next two to three years. It also discusses some reasons why enterprise IT has to pay attention to this trend, and not dismiss it as “ASP revisited.”

Traditional ERP and CRM

SaaS is starting to enjoy considerable success in the marketplace. Since 2001, has picked up more than 30,000 customers, and the subscriber base is growing by 50 percent to 75 percent per year. And, Dave Duffield’s SaaS-based replacement for ERP applications such as those delivered by his first-born company, Peoplesoft, is only partially deployed but already has its first subscribers.

Many in the industry are starting to scratch their heads and wonder, “What is different? Is this just ASP re-packaged?” The good news is, the answer is no. SaaS is an entirely new model engineered specifically for multi-tenant delivery and economies of scale.

Although the idea behind ASP was likely ahead of its time, it had multiple disadvantages from the start. ASP deployments attempted to repurpose software designed for a single enterprise to a hosted, multi-customer paradigm. The tiered, multi-server architectures characteristic of ERP and CRM vendors required servers for databases, applications, and Web connections to interact in a way that delivered a business application to a single company. Such architecture didn’t lend itself to the replication required to support multiple companies.

Adapting such applications to multiple customers proved costly. Duplicating tiered applications added hardware and power costs. And multiple software versions added expense, as well. Supporting such deployments became complex, and therefore expensive. The result was an expensive delivery model where customers benefited primarily from predictable cost, not necessarily from cost savings.

If traditional software vendors want a piece of the action, then they will have to overcome these same problems. Multi-tenant deployment requires significant modifications to software architecture and code. ERP and CRM vendors, for example, already spend an estimated one-third of their research and development (R & D) budgets on supporting multiple product versions and another third on supporting multiple platforms. This leaves a relatively small percentage for innovation and for re-architecting for SaaS.

Seriously pursuing the SaaS market would also require significant changes in both revenue models and corporate culture. Oracle and SAP have built enormous sales organizations around selling licenses and support. Refocusing sales executives on selling subscriptions would be a “hard sell”. Refocusing corporate culture from selling software to providing a service is another aspect of the SaaS model that would be problematic as well.

What makes SaaS different?
Today’s SaaS vendors have capitalized on the lessons learned by their predecessors and are addressing them with innovations that yield big advantages. They include:

One version, one platform: While legacy vendors still focus the majority of their resources on supporting multiple versions and platforms, supports only one version running on one platform. One hundred percent of Salesforce’s resources can be devoted to innovation, resulting in a product evolution that is very agile and much more in touch with “real-life” IT requirements.

Next page: An area of savings…