The current economic downturn has pushed the Software as a Service (SaaS) hype machine into overdrive. For all the turbocharged hype, SaaS does not solve the core business problem it purports to address – rising costs of information technology. Information Technology costs comprise several elements: hardware, software licenses, implementation services, programming services, and ongoing software maintenance fees. The idea behind SaaS is to bundle select cost elements into a “low” monthly rental fee. Implementation and programming services are provided on a fee for service basis by the SaaS vendors. Over time (typical information systems are used for more than 10 years) this low fee adds up to more than the component elements would – unlike a purchase the fee never goes down once the asset(s) is paid off. A better approach is to address the cost elements of traditional information technology delivery in detail.
Hardware costs follow Moore’s Law and decrease over time. Hardware is not a major cost factor in a modern information technology project. Overhead associated with server farms can be very expensive (and is a big contributor to the costs of delivering SaaS), but these costs are easily avoidable by investing in an integrated business computing platform such as the IBM i.
Software licenses are a significant upfront expense. Poorly designed software licenses add considerable expenses through the life of an information system (think costs for additional seats, upgrades, cpu transfers). These late life cycle license costs are unpredictable, who knows what a vendor will think up next? Newer license forms, such as the Omni-License, provide cost certainty that may be budgeted for and financed over several years. The Omni-License when leased for 3 or 5 years compares favorably SaaS rental fees with the bonus that payments end well before the information system needs to be replaced.
Software maintenance fees typically depend on a monopoly approach towards customers, a feature shared with SaaS rental fees. Vendors assume that the costs to change a system before its allotted 10 years is high enough that customers have no ability escape any price increases imposed upon them. Please note that the Omni-License does not require continued maintenance, but earns customer maintenance renewal by providing new business value throughout its 5 year term. Ongoing maintenance is offered at a significantly lower price per year than the initial Omni-License.
Taken together, the costs as bundled in SaaS are greater than for a traditional licensing approach, and much greater than for an Omni-License. Software as a Service fails to address the fundamental elements of increased information technology cost.
What about the truly expensive part of information technology, the implementation and programming services? SaaS has nothing new to offer here. There must be a creative new approach to dramatically reduce all the services attached to software. This is where Service as Software comes in.
Service as Software contains two main themes. First the software products must be designed to reduce or eliminate the need for services in the first place. Second remaining service requirements must be packaged for on-demand delivery as an “all you can eat” subscription. Benefits include reduced quantity of services required, a reduced cost of services consumed, and timely delivery of services needed.
Designing software to reduce the need for post sale services is a choice that few vendors make. If you consider your customer base a monopoly to be exploited, and one third of your revenues come from selling services into that monopoly what incentive do you have to reduce the quantity of services required? None. It is easy to tell if a vendor chose to design their software to reduce required services – look at their income statement and see how much they depend on services. HarrisData service revenues today are about 6% of the total. We expect this number to be lower in the future. How did HarrisData reduce service revenues to such a low percentage? By design and by working with our customers. Highlights include exit points for customer program extensions to buffer software upgrades and fixes and HarrisData Widgets for user customizable screens.
Packaged service offerings address the remainder. For example one major component of implementation costs is end user training. End user training is provided as a last step before going live. For customers with multiple locations to be trained, end user training may be repeated at each location. Post implementation, end user training is repeated as employee turnover impacts the user base. Additional end user training is required for specific topics tied to infrequent activities (e.g. year end payroll processing). Over the life of an information system end user training is a big need. How then can the cost be brought under control? HarrisData is developing a library of end user training chapters to be provided on demand to our customers. The library will include suggested (or custom) syllabus complete with course tracking by individual user. Subscriptions to the library will allow customers to ensure that their users are trained as needed for predictable (and affordable) cost.
We are open to other ideas for packaging post sale services ala Service as Software and welcome your suggestions.