Enterprise software industry practitioners and observers are continuously bombarded with messages about impending world domination by software-as-a-service (SaaS). The view of the future of SaaS is unfailingly positive, irrespective of the enterprise software market under discussion. But the term “SaaS” is rarely defined with any rigor and frequently conflated with “Cloud .”
It has been more than a decade since the launch of Salesforce.com, yet it is hard to find an equivalent success story beyond a handful of major markets – CRM, human capital management (HCM), and collaboration. Across many enterprise software markets, such as enterprise resource planning (ERP), supply chain management (SCM), business process management (BPM), and business intelligence (BI), SaaS remains a marginal force at best. This warrants a serious evaluation of where SaaS works and where it does not, and why.
SaaS only offers an advantage over the status quo when certain conditions hold
It would be appropriate to start such an evaluation with a definition of SaaS. For an application to qualify as SaaS it must meet three conditions: the architecture should be multi-tenant, the application should be hosted off-premise in a data center not owned or managed by the end-user company, and licensing should be term-based.
In an attempt to analyse the issue based on first principles, it would be useful to consider the total cost of ownership (TCO) for software implementation. TCO for software implementation comprises: application licensing; hardware; infrastructure software licensing; customisation; integration; business process re-engineering; and training.
SaaS impacts just three areas: hardware, software infrastructure (such as database software), and licensing. The rest are functions of organisational complexity and are not dependent on the business model or the delivery model. Of the areas impacted by SaaS, it is not entirely clear if licensing costs are always lower for SaaS – term licenses are hardly new in the on-premise software business. Monthly license payments are also largely a myth, in Ovum’s experience. Annual subscriptions are much more common, which are very similar to term licensing schemes.
It is logical to assume that SaaS is the cheaper option when licensing, hardware, and software infrastructure account for a high percentage of TCO. This tends to be the case when customization, integration, business process re-engineering, and training costs are relatively low. Such elements are likely to register a collectively low value if:
- a standard software solution bought off-the-shelf is likely to suffice (so customisation and process re-engineering costs are low),
- the user base is from a single department (so integration costs are likely to be low),
- users are already trained in the tool, for example if it has existed for some time, such as CRM, or is used outside of work, such as collaboration (so training costs are low).
- Also, when a standard solution does suffice, a multi-tenant architecture is possible. This creates an important cost advantage: low application maintenance cost.
In addition, the SaaS low-touch, low-cost sales model is only possible when a standard solution is likely to suffice, as it reduces the deal value and makes the proposition easy to assess, and when the user community is from a single department, as it reduces the number of hoops through which the sales team must jump.
Ovum maintains a list of top vendors in every major enterprise software market, defined as the largest vendors which collectively account for 50-80% of each market. As part of a research study on the importance of SaaS, Ovum calculated the percentage of top vendors in each market which provide a SaaS solution.
For the CRM, HCM, and collaboration markets, 40%, 50%, and 64% of the top vendors, respectively, provide a SaaS solution. For BI and BPM, 20% and 26% do, respectively. For SCM, less than 5% do. Likewise, by Ovum’s estimates, only the CRM, HCM, and collaboration markets are served by SaaS vendors which feature in the top 14 publicly traded SaaS companies.
Whatever metric we choose, the results are conclusive: SaaS is a lot more successful in enterprise software markets in which standardized solutions suffice. Maintaining contacts and sales records, managing the employee appraisal cycle, and encouraging groups to talk in a Facebook-like environment are tasks that differ far less across companies and industry sectors than managing production processes, reporting production performance, and managing the supply chain.
Recommendations for enterprises and vendors
Enterprises must assess the words “Cloud ” and “SaaS” very carefully. If they are associated with solutions that are unsuitable for SaaS, it is unlikely that the purported benefits of SaaS will materialise. A solution could still be a good candidate because of its functional superiority, but the much-hyped benefits of SaaS would probably not be realised. Also, an application could be based on a single-tenant architecture, hosted off-premise, and referred to as “Cloud ” – but Cloud does not equal SaaS.
For vendors, it is important to disaggregate SaaS into the following components:
- SaaS is a way of providing low upfront costs. But this can also be provided through a term license, which does not necessitate a SaaS model or the use of “SaaS” as a marketing term.
- SaaS is a way of setting up application access quickly and providing low time-to-value. But there are situations in which the need for low time-to-value can be addressed with an appliance model or a low-deployment-time methodology (such as templates). Use of the term “SaaS” in such circumstances is quite inappropriate and likely to be viewed with scepticism. The most inappropriate term that Ovum has heard is “SaaS in a box.”
- SaaS is a way of enabling application access from anywhere. For all but the largest of enterprises and the most resourceful of IT organisations, deploying a high-usability web-accessible application is much tougher than simply buying subscriptions. It does not require a multi-tenant infrastructure, and appropriating the term “SaaS” for the ancient Application Service Provider (ASP) model is likely to be viewed with scepticism.
Source: BusinessCloud 9, Somak Roy. 20th February 2012
