An independent paper produced by Diversity Analysis with the support of Intacct 1 Copyright 2010 - Diversity Analysis Contents Introduction ................................................................................................................................2 Comparison Chart: Software Delivery Models ....................................................................3 An Overview of the Tree Major Software Delivery Models ..........................................4 On-Premises Software Deployment ............................................................................................... 4 Hosted Software Deployment ......................................................................................................... 5 Software as a Service (SaaS) .............................................................................................................. 6 Te Pros and Cons at a Glance ................................................................................................7 IT and Operational Requirements ..........................................................................................8 Software Procurement ....................................................................................................................... 8 Implementation ................................................................................................................................... 9 IT Operations and Management ...................................................................................................10 Security ...............................................................................................................................................11 Upgrades .............................................................................................................................................12 Business Continuity ..........................................................................................................................13 Ease of Customization .....................................................................................................................14 Service and Support ............................................................................................................... 15 Total Cost of Ownership ....................................................................................................... 16 Conclusion: Software Orientation, Service Orientation ................................................. 18 About Intacct .......................................................................................................................... 19 About Diversity Analysis ...................................................................................................... 20 About the Authors ................................................................................................................. 21 An independent paper produced by Diversity Analysis with the support of Intacct 2 Copyright 2010 - Diversity Analysis Introduction Relentless innovation in the way businesses apply computing power to automate processes has translated into tremendous gains in just a few short decades. Te unprecedented benets we enjoy today are virtually taken for granted - greater velocity, productivity and accuracy along with lower overall costs and greater operational visibility and insight. A chief contributor to these improvements is the delivery model that businesses adopt for consuming and using computing resources. Te computing industry has evolved from on-premises software implementations to o-site hosting (and its oshoot, the application service provider or ASP model) to Software as a Service (SaaS). Despite signicant dierences among these models, a wide range of misperceptions and myths persist about each models relative merits and potential drawbacks. Some even believe that this market confusion caused by mislabeling and proliferating jargon is a deliberate industry strategy to jockey for a marketing advantage. Te following white paper explains, contrasts and compares the three major software delivery models and provides buyers with useful criteria to evaluate software oerings and determine the suitability of the approaches to their unique environment and needs.
An independent paper produced by Diversity Analysis with the support of Intacct 3 Copyright 2010 - Diversity Analysis Comparison Chart: Software Delivery Models On-premises software Hosted software Software as a Service Application Development Developed as one- size-ts-all model Borrows on-premises software and modies it for online delivery Developed from the ground up for online delivery Application Deployment Installed on the customers own hardware Installed on vendors hardware delivered via the Internet Installed on vendors hardware delivered via the Internet Implementation Lengthy implementation time Lengthy implementation time Faster implementation timeframes Customization Possible but expensive and time-consuming Limited customization tends to be very expensive Customization limited but an inherent part of the application Application Design Monolithic client/ server designs Monolithic with a simple HTML interface Designed for the Web environment and generally lighter weight Upgrades 12+ months Depends on developers upgrade cycle Generally monthly or even more frequently Integration Can be dicult and expensive Can be dicult and expensive Readily available via application programming interfaces (APIs) IT Support Generally provided by the customer Generally provided by the customer, with some vendor support Generally included in the package from vendor Multi-Tenancy Not applicable Each customer requires an individual instance Applications are designed to be multi-tenant Hardware Requirements Requires a specic operating system environment Can require specic applications and operating environments Delivered via a Web browser so generally operating system- and browser-agnostic Accountability Generally single- vendor so support path is well-dened Hosting provider and software developer are two dierent organizations so accountability is complex One vendor provides end-to-end solution so accountability is inherent An independent paper produced by Diversity Analysis with the support of Intacct 4 Copyright 2010 - Diversity Analysis An Overview of the Tree Major Software Delivery Models Perhaps one of the most important and fundamental steps we can take to understand the dierence between the three major software delivery models is to properly dene each model. Only then can we trace their evolution and understand their merits and drawbacks.
On-Premises Software Deployment Before the widespread availability, aordability and adoption of networks particularly the Internet on-premises software deployments were virtually the only choice for businesses. Some examples include large-scale enterprise resource planning (ERP) and customer relationship management (CRM) systems as well as single-user programs such as QuickBooks or Microsoft Oce. As its name implies, an on-premises implementation is software thats installed and operated on computer(s) located on the premises of the software licensee, rather than at a remote facility. Tis model largely dened and drove the rst generation of business computing. However, on-premises software is limited in its ability to support remote access to computing services. Customizations if allowed can be dicult and expensive. Software vendors also make signicant investments in legacy code that tend to work poorly in o-premises congurations.
An independent paper produced by Diversity Analysis with the support of Intacct 5 Copyright 2010 - Diversity Analysis Hosted Software Deployment Te increased penetration of the Internet proved to be an unstoppable inuence on business software. In the 1990s, the combined forces of IT outsourcing and widespread Internet adoption led many vendors to develop on-premises-style commercial software but host it o-site and license it to many dierent customers. In a hosted deployment, the software physically resides at a third-party data center the hosting site where it is operationally managed by a service provider on behalf of the licensee. One o-shoot of this approach is the application service provider (ASP) model where the vendor provides additional domain expertise. ASPs specialize in software categories or even specic software vendors, helping to customize implementations for each customer. Regardless of whether it is through an operational hosting provider or ASP, users access these remote computing resources over a network, typically the Internet. Each client company accesses its own dedicated instance of the application that is running at the remote data center which creates signicant overhead costs. Te compelling value proposition is that the cost to own and operate the software is far less if entrusted to an expert organization with the infrastructure and expertise solely dedicated to delivering that resource to the customer. Much of the IT overhead cost is transferred to the hosting provider no longer does the organization need to build, sta, or maintain its own data center or support desk. Unfortunately, much of the optimism surrounding the hosted software- delivery model was unfounded. Flawed business models particularly for ASPs - combined with integration challenges have left this market segment in a state of stagnation. Many hosting providers are not the actual developers of the software program, leaving them largely at the mercy of the developer in terms of development timetables/roadmaps and features. Whats more, most An independent paper produced by Diversity Analysis with the support of Intacct 6 Copyright 2010 - Diversity Analysis hosting providers are typically better-versed in hardware and networking issues and unused to dealing with software delivery and customization as well as end-user customer support challenges. Software as a Service (SaaS) Aiming to combine the best characteristics of both on-premises software and outsourced hosting, the Software as a Service (SaaS) model (sometimes referred to as cloud computing) has emerged as an increasingly compelling alternative for many licensees. With SaaS, the application developer delivers shared, scalable services that the customer accesses over the Internet using a Web browser or mobile device. Teres no need for the customer to own, license, understand, manage, or control the underlying hardware, software, or data/networking infrastructure that supports the service. SaaS is typically delivered on a term-based subscription basis, eliminating the need for upfront software licensing fees or major hardware purchases. Te result is cost-eective, anytime/anywhere managed access thats increasingly attracting a growing number of enterprises and small/medium-sized businesses. One of the key tenets of SaaS applications is a concept called multi-tenancy. In SaaS deployments all customers simultaneously access the same single instance of the application in the remote data center. Tis means the vendor can amortize a world-class infrastructure across multiple customers while reducing costs. Whats more, well-designed SaaS applications allow for instant provisioning and upward/downward scalability, parameter-driven customization and conguration and regular upgrades (instead of the daunting customization- breaking upgrades required with on-premises or hosted software implementations). So which model is best for your organization? As you conduct your evaluation, consider the following metrics and factors.
An independent paper produced by Diversity Analysis with the support of Intacct 7 Copyright 2010 - Diversity Analysis Te Pros and Cons at a Glance On-premises Software Pros: Greater end-to-end control of your software Maintain IP within the organization Implement signicant customization Cons: Generally higher cost Requires you to build and maintain your own IT infrastructure Must provide technical support, upgrades, and version control Upgrades may break any previous customizations Integration with third-party applications can be dicult/expensive Hosted/ASP Software Pros: Pushes infrastructure costs to the provider Potentially lower cost than on-premises deployment Cons: Very limited customization Upgrades may break any previous customizations Support and service is typically an expensive extra charge Integration with third-party applications can be dicult/expensive Single tenancy limits the cost-reduction opportunities Customers have some security concerns Generally poor usability Software as a Service Pros: Core architecture makes it the most economical choice Ability to provision almost instantly Pushes infrastructure, service, and support costs to the vendor Integration is quick and easy using application programming interfaces (APIs) Customization available Upgrades are automatically applied to all users - all customers work with the same application Encourages alignment of people/process/system Cons: Customers have some security concerns An independent paper produced by Diversity Analysis with the support of Intacct 8 Copyright 2010 - Diversity Analysis IT and Operational Requirements Some of the more important considerations when choosing a software deployment model involve the software and hardware required and the people and skill sets required to properly operate and maintain that IT infrastructure. Te following issues are key to any decision. Software Procurement Simply put, who procures the product and in what manner? With on- premises or hosted software, you will need to enter into a termed license and also purchase a maintenance contract simply to take physical possession of the software. And with on-premises software, the IT department must have a leadership role in the evaluation/selection process because they will have to manage the software as well as that softwares potential impact on the rest of the corporate IT portfolio. By contrast, with SaaS, the business user can drive the procurement process consulting appropriately with IT colleagues. Whats more, since SaaS is procured through a monthly, quarterly or annual subscription fee, there are no upfront licensing fees (which often reach into six or seven gures) and theres a lower approval threshold and risk prole. An independent paper produced by Diversity Analysis with the support of Intacct 9 Copyright 2010 - Diversity Analysis Implementation Who is responsible for implementing the chosen solution? For on-premises deployments, many companies nd that they must implement the software themselves (a process that can take months and sometimes years) or pay large dollars for expert consultants and implementation partners. Hosting providers handle much of the implementation headaches theyre experts, after all but they are still deploying a dedicated (sometimes customized) instance of the software, which can take almost as long as on-premises software deployments. For SaaS implementations, either the vendor or one of the vendors pre-approved implementation partners is responsible for rapidly provisioning access to the online application. Even with customizations, complex SaaS implementations can be completed in just days or weeks. Naturally, a lengthy implementation cycle that stretches for many months means theres a far greater nancial and operational risk to the customer. From software licensing to hardware to labor and consulting fees, signicant costs are incurred from day one. But process and productivity improvements and the payback they represent can often be elusive, stretching years out into the future. By contrast, SaaS subscriptions require no upfront licensing fees and no lengthy implementation cycles and no attendant costs, translating into a far lower level of project risk.
An independent paper produced by Diversity Analysis with the support of Intacct 10 Copyright 2010 - Diversity Analysis IT Operations and Management Todays enterprise IT sta is already bearing a signicant burden, so any new addition to the application portfolio must survive intense scrutiny and analysis before winning their approval. Major enterprise applications entail procurement, deployment, operations and maintenance cycles for software and IT infrastructure tasks that fall to the licensee in the on-premises model or the hosting provider in the dedicated-instance hosted-software model. With SaaS, those costs shift to the commercial software developer. Te customer need only sign a subscription contract and the application service is accessible in much the same manner as a basic utility service: on-demand and scalable as needed. SaaS vendors achieve this by sharing their world-class infrastructure across thousands of customers. Whats more, as the developers and full-time experts in providing this application service, the SaaS vendor is uniquely qualied to optimize the users experience with respect to performance, availability and security. Teres no internal IT learning curve to climb.
An independent paper produced by Diversity Analysis with the support of Intacct 11 Copyright 2010 - Diversity Analysis Security Te ability of on-premises deployments to provide complete control over the application and infrastructure is initially quite appealing. Many organizations have resisted hosted or SaaS models because they dont feel comfortable storing and managing sensitive data in locations not under their direct control. However, putting security responsibilities on an internal team provides drawbacks. For example, the cost of proper information security measures can be prohibitive. Furthermore, on-premises customers must rely on the software vendor to release appropriate security patches to close any vulnerabilities in the application that may emerge. With third-party application hosting, security is outsourced, cutting costs considerably. However, the eectiveness of that security depends on the size and capabilities of the service provider. Te single-tenant architecture localizes any attack to a single customer and reduces the chances of any security attack aecting all customers. However, this can lead to complacency and move focus away from the security of their service. As with on-premises deployments, third-party hosters also passively rely on the software provider for security patches, opening them to the same disadvantages. SaaS skeptics suggest that a lack of direct control over SaaS applications and data could result in a security nightmare for a business. However, this is a myth based on the erroneous assumption that SaaS providers cannot employ security professionals with the same strong expertise that many enterprises do. Te reality is, a security breach for a SaaS company would aect all of its customers, which provides an incredible incentive to provide ironclad security measures. Teir business is predicated on the ability to secure their infrastructure, applications data and operations. Whats more, the SaaS model has a distinct advantage over on-premises and outsourced hosting models when it comes to patching application vulnerabilities. Since the SaaS provider is also the application developer, it can instantly patch application vulnerabilities. An independent paper produced by Diversity Analysis with the support of Intacct 12 Copyright 2010 - Diversity Analysis Upgrades As any enterprise IT professional will tell you, upgrades are a never-ending nightmare in many organizations today. Just deploying the application (or applying a previous upgrade) can take months. In many instances, simply achieving a stable, reliable production environment is an extremely challenging task. And when youre done? Time to do it all over again. On-premises software means internal enterprise IT teams can be consumed with an endless cycle of upgrades to business critical applications. Whats more, the lengthy timeframes to upgrade can jeopardize service contracts or potentially introduce instabilities and incompatibilities with other operating systems, peripherals, applications and tools. Te alternative, of course, is to hire an expensive team of consultants to hopefully expedite the process and lower the risk. For third-party application hosters, experience can be a major advantage. After upgrading a customer for the 20th time, theyre likely very skilled. However, that takes time and consumes lots of resources that customers may need for other matters. Whats more, upgrades can wreak havoc on both on-premises and hosted software deployments. New versions often break previously developed/applied customizations to commercial code. By contrast, SaaS oers signicant advantages in the upgrade process. First and foremost: there are no signicant tasks for the customer. Some well- communicated advance notice likely mitigates any issues of disruption or preparation. Te SaaS vendor is the application developer, so there are no learning curve issues. And in a multi-tenancy model, the upgrade only needs to be applied once to reach all customers. Teres no need to wait your turn while waiting for a hosting provider to wend its way through a backlog of customers seeking the upgrade. Whats more, since all customizations are already built into the commercial code, the upgrade has no disruptive impact on the production system. For a SaaS customer, an upgrade is a non-event, simply appearing one day without any intervention required. An independent paper produced by Diversity Analysis with the support of Intacct 13 Copyright 2010 - Diversity Analysis Business Continuity As the operator of the software, the on-premises IT team or outsourced hosting provider must perform all of the routine but critical backup and o-site data storage tasks to ensure business continuity in the event of natural disaster, re, civil unrest, or other interruption. Teres no substitute for the business continuity plan and omitting it is not an option. And that means: wasted time, expense and overhead devoted to non-strategic activities especially for on-premises deployments. By contrast, a SaaS implementation alleviates the need for separate business continuity eorts related to that SaaS-deployed application. As part of the regular application subscription, the vendor backs up all customer data (according to pre-dened service level agreement parameters) o-site and deploys separate data centers and plans to provide uninterrupted service during any potential disasters or major unplanned downtime events. An independent paper produced by Diversity Analysis with the support of Intacct 14 Copyright 2010 - Diversity Analysis Ease of Customization Regardless of deployment model, when a software solution doesnt meet all your needs out of the box, an important factor is how dicult is it to congure, customize, or tailor the application to your unique requirements? On-premises software oers excellent opportunities to build a customized solution and provides the highest level of customer control. However, highly customizable environments are expensive, time-consuming and not without risks. For instance, you might need to make changes to database tables, write new code in Visual Basic, or spread your customizations into many dierent areas of commercial code that was written by the vendor. (Of course, for lower-end commercial software, such as QuickBooks, there may not be any customization opportunities at all.) Whats more, when commercial upgrades are applied, your expensive customizations may break and require lengthy rework and retrotting. Regression testing of your customizations if its done at all might overlook subtle issues, incompatibilities and security vulnerabilities that youve unwittingly introduced. Its a similar story with third-party-hosted, dedicated instances of the application. One of the major shortcomings of outsourced application hosting is the lack of domain expertise, making integration with other systems a major headache. One of the biggest myths about SaaS is that it lacks customization and integration options. Unlike what naysayers want you to believe, SaaS applications are rich with customization and integration features that are easy to access and congure. Although most SaaS applications oer customization from a branding point of view, many enterprise-level SaaS applications are now also oering the option to customize based on functional needs. Te newest generation of SaaS applications actually oer customization capabilities that exceed whats possible with traditional on- premises software. Most importantly, SaaS customizations are preserved through the upgrade process, so the upgrade becomes a value-added non- event for the customer. An independent paper produced by Diversity Analysis with the support of Intacct 15 Copyright 2010 - Diversity Analysis Service and Support Te issues of service and support are somewhat easier to overlook, but their importance cannot be overstated. When you operate an on-premises software solution, you assume signicant burdens and costs. You need to expand your existing (or even create a new) support desk for the application service. Tat means recruiting, training and retaining talented experts and delivering support and coverage as business needs dictate to support users. Tere are IT-facing issues as well. Te on-premises customer must ensure the infrastructure devoted to the application is running at peak performance with acceptable levels of reliability, availability, security and performance. Ten, of course, when issues arise, it can be daunting to track down the source of the issue and hold the appropriate vendor accountable. Get ready for a lot of nger-pointing. Collectively, thats a tall order, which is why many outsource these tasks to hosting providers. However, many hosted- application arrangements are characterized by confusion over who provides the support: the hosting provider, the application developer, or the customer. In some models, the hosting provider provides direct support to end users. In other instances, the hosting provider only trains the customers IT help desk, which then directly handles level-one support calls from users. Clarity on these matters is essential. In the SaaS application-delivery model, its simple. Tere is a single vendor providing the entire service who is responsible and accountable for both the product and service/support issues. An independent paper produced by Diversity Analysis with the support of Intacct 16 Copyright 2010 - Diversity Analysis Total Cost of Ownership Any comparison of software deployment models must address the critical issue of cost. In this paper, we approach this from a total cost of ownership (TCO) perspective that encompasses everything from licensing or subscription fees to labor costs and infrastructure investments. Even before the current economic downturn, businesses scrambled to reduce their operational costs. Not only are they looking to cut overall costs, but many are seeking to move costs out of capital budgets and into operating budgets. Te drivers for this are simple - capital expenditures tend to have a long and complex approval process and create a disconnect between the revenue generated by a nancial outlay, and the outlay itself. Moving expenditures to the operating line more closely aligns them with revenue generation, eases the approvals process and avoids the sunk cost issue where organizations are reluctant to move away from projects with signicant historical capital expenditure - regardless of their eectiveness. On-premises is the most expensive of the three models, presenting a very high TCO. Te infrastructure costs for running the data center, licensing fees, upgrade and maintenance fees and labor costs result in extensive capital and operating expenses. Te restrictions and barriers presented by traditional software applications combined with the responsibilities arising from deployment and management of the applications inside the data center often makes on-premises applications prohibitively expensive with uncertain ROI and very long payback periods. Other factors contributing to the runaway TCO of on-premises applications include implementation costs and internal support costs. In a challenging economic climate, the on-premises model may well turn out to be a business handicap for any organization. An independent paper produced by Diversity Analysis with the support of Intacct 17 Copyright 2010 - Diversity Analysis Although outsourced hosting is less expensive than the on-premises model, it still presents higher costs that SaaS because of the single tenancy delivery model and the signicant licensing costs. Tere are other components to TCO like the cost of application upgrades, additional support costs and more. Compared to the SaaS model, the outsourced hosting model has proven to be very cost inecient. SaaS is the most cost-eective model with a very low TCO. First, the SaaS vendor assumes more responsibility than hosting service providers - demanding no capital expenditures from the customer. And since the SaaS vendor is also the actual application developer, licensing costs are also not a factor. SaaS vendors utilize a multi-tenant architecture to save on the costs of oering the application as a service. Tis cost savings is passed on to customers resulting in a lower TCO. Whats more, the absence of any upgrade, implementation, or maintenance fees accelerates the payback of the SaaS approach. Te cost components of SaaS are for subscriptions and training. Compared to the huge upfront licensing and training fees for traditional on-premises software, the cost of SaaS is much lower.
An independent paper produced by Diversity Analysis with the support of Intacct 18 Copyright 2010 - Diversity Analysis Conclusion: Software Orientation, Service Orientation Given the numerous operational, functional and nancial advantages, SaaS deployments are gaining increased attention from enterprises seeking smarter, more ecient ways to access and deploy critical computing resources inside and outside their organizations. With its software orientation and service commitment, SaaS in many ways represents a best of both worlds perspective, combining the control and innovation of on-premises software with the speed and convenience of the outsourced hosting models all with rapid innovation and lower cost. As this innovative software-delivery model gains further momentum, software evaluators would be well-advised to give SaaS careful consideration. An independent paper produced by Diversity Analysis with the support of Intacct 19 Copyright 2010 - Diversity Analysis About Intacct Intacct is the market and technology leader in cloud computing nancial management and accounting applications for small and midsized businesses, and the preferred nancial applications for AICPA business solutions. Intacct applications are used by thousands of businesses from startups to public companies and are designed to improve company performance and make nance more productive. Te Intacct system includes project accounting, contract management, revenue recognition, inventory, purchasing, vendor management, nancial consolidation and nancial reporting applications, all delivered over the Internet via Software as a Service (SaaS). Intacct is headquartered in San Jose, Calif. For more information, please visit www. intacct.com or call 877-437-7765. An independent paper produced by Diversity Analysis with the support of Intacct 20 Copyright 2010 - Diversity Analysis About Diversity Analysis Diversity Analysis is a broad spectrum consultancy specialising in SaaS, Cloud Computing and business strategy. Our research focuses on the trends in these areas with greater emphasis on technology, business strategies, mergers and acquisitions. Te extensive experience of our analysts in the eld and our closer interactions with both vendors and users of these technologies puts us in an unique position to understand their perspectives perfectly and, also, to oer our analysis to match their needs. Our Analysts take a deep dive into the latest technological developments in the above mentioned areas. Tis, in turn, helps our clients stay ahead of the competition by taking advantage of these newer technologies and, also, by understanding any pitfalls they have to avoid. Our Oerings: We oer both analysis and consultancy in the areas related to SaaS and Cloud Computing. Our focus is on technology, business strategy, mergers and acquisitions. Our methodology is structured as follows: Research Alerts Research Briengs Whitepapers Case Studies We also participate in various conferences and are available for vendor briengs through Telephone and/or Voice Over IP. An independent paper produced by Diversity Analysis with the support of Intacct 21 Copyright 2010 - Diversity Analysis About the Authors Ben Kepes Ben is the founder and managing director of Diversity Limited, a consultancy specializing in Cloud Computing/SaaS, Collaboration, Business strategy and user-centric design. He is also editor of Diversity Blog, a specialist Cloud Computing blog. More information on Ben and Diversity Limited can be found at http://diversityanalysis.com/ Krishnan Subramanian Krish is an entrepreneur, open source researcher and evangelist, cloud computing analyst and consultant, ex-physicist, and one of the industrys most prolic bloggers. He is part of the management team in two companies in India and sits on the board of a few privately owned companies. More information about Krish can be found at http://www. krishworld.com/.