"Application service provider" as a term is dead. It carries significant negative connotations, especially within...
the financial community, and it will never return. The problem with ASPs is in the business model, not in the hosted applications. Indeed, applications are the fuel feeding the fire of offshore outsourcing and other high-growth outsourcing sectors.
Currently, IT managers are frustrated with the lack of commoditization for application services. Pricing for other outsourcing services (servers, storage, desktop, etc.) have standard and known market ranges that allow for market comparison. Applications lack any similar structure. Thus, while an IT manager can compare labor rates within the market, the same manager is uncertain as to whether a project requires 1,000 developer hours or 10,000 developer hours. Application services do not have common building blocks.
Vendors are taking various market approaches to create better and faster predictability in application services:
- Keane: Keane has developed a rapid project assessment methodology that identifies common components of an application project and increases the accuracy and speed of estimates. Since 2000, Keane has continuously reduced average quote time despite the lack of a "silver bullet" for estimation.
- Oracle: Oracle's strategy focuses on standardization (similar to IBM's On Demand philosophy). By standardizing clients' operational environments, Oracle Outsourcing reduces complexity and is able to increase predictability of costs. Despite its success with this model, Oracle indicates that more than 70% of customers request customization outside of the standard pricing model, adding costs to the standard pricing model.
- IBM: IBM is creating application components within the context of its overall On Demand initiative. Application components will transcend specific application environments (e.g., .Net, Java), specific applications (.e.g., Oracle, SAP), or legacy versus packaged (e.g., COBOL). Application components not only standardize the technology elements, but also create a common pricing structure that simplifies client understanding of pricing drivers. IBM is currently testing components, and its offerings will emerge in 2H04.
We believe these various concepts of standardization and components will converge into a common industry pricing model by 2006/07. Until that time, however, IT organizations must work within the differing pricing structures of industry providers. Moreover, assessing the size and scope of application work will remain a difficult process, requiring either bids from multiple vendors for comparison or internal skills to scope the project and challenge or debate vendor estimates.
Concomitant with strategies to streamline and commoditize application services are other market dynamics that would obviate the application technology discussion completely:
- Per-user pricing:
Pricing per user requires a sophisticated understanding of the application environment, including infrastructure, bandwidth, performance, and various user profiles (i.e., named, concurrent, light). User-based pricing was a key premise of the original ASPs (circa 2000/01). Although many ASPs offered this type of pricing, it was done without a thorough understanding of the underlying costs. As these ASPs lost money due to overaggressive pricing, the vast majority went out of business. Now, several years later, leading vendors are just beginning to have adequate knowledge to offer user-based pricing. Pricing per user is currently feasible only for packaged applications. Legacy or custom development environments can rarely adopt this simplified pricing option.
- Business process outsourcing (BPO) enabler:
Some vendors are completely bypassing the application to focus on providing a business solution (e.g., invoice processing). The client pays for business transactions and, in fact, may not even know what applications support or enable the business process. This option is feasible only in highly commoditized or standardized business processes, where the client pays for a type of result or output (e.g., check processing in the banking industry). Most applications and business operations are not mature enough to use this option.
Although the outsourcing industry will introduce some improvements into application services, these services will continue to lack dramatic innovations (i.e., silver bullets) for several years. IT organizations should assess the following vendor capabilities:
- Hosting experience with companies of similar size, scope, and strategic direction:
This includes future needs for e-commerce, Web integration, Web enablement of legacy applications, or supply chain initiatives.
- Experience in specific (or related) vertical markets: This includes understanding of industry trends, providing business information services or connections/relationships within the industry.
- Long-standing partnerships that provide needed capabilities:
It is not important that the company provide all services, but it must avoid "new" relationships, since it takes years to resolve relational issues.
- Ability to convert existing data stores or legacy data into the new application environment:
This includes assessment of the limitations of "vanilla" application capabilities and of the options to determine viability. ERP vendors should not be the primary vendor if more than 10% customization is required.
Bottom Line: Application solutions are currently the least commoditized of the outsourcing services. Although some customization will remain necessary, building blocks and common pricing structures will improve through 2008/09.
Business Impact: New pricing structures for application services will improve the measurement and alignment of IT services with business objectives and will also simplify the cost/benefit analysis of IT on business results.
Click here for more business and IT research from META Group
Copyright ©2004 META Group Inc.