29 Oct 2015
Why implementing SaaS isn’t easy – and how to make it easier
There is a growing trend among organizations to implement software-as-a-service (SaaS) applications as a way to update or improve their business software. Aberdeen have reported that SaaS is the most common form of cloud usage. When businesses initially evaluate SaaS software it seems cheaper, easier to maintain, and generally doesn’t involve the headaches of running on-premise software in-house. There’s a catch, though. Integrating SaaS software with existing legacy systems, as well as other SaaS applications, can be a struggle.
ERP moves to a postmodern period
There was a time not long ago when companies that wanted new software had to go through the process of paying for the software upfront, then installing and testing it, not to mention supporting, maintaining and later upgrading it. A common preferred route was adding a new module or sub-system to the company’s existing enterprise software (e.g., ERP, CRM); it was preferred because the new add-on would be already integrated with the rest of the business. However, it could often be a compromise solution, with the new add-on being “not quite right, but good enough”, and it was quite possible for implementations to take months. This practice of relying on a vertically integrated stack of applications is being challenged by SaaS applications which offer the promise of greater functionality, decreased implementation time, and cost-effectiveness of cloud computing. It’s what Gartner calls the move to a postmodern ERP strategy.
The challenges of SaaS
The downside that businesses are learning about SaaS is that it can create silos of information. Getting that information shared between systems becomes an integration issue. For the first one or two SaaS applications the point-to-point integration work is relatively straightforward. But when demands for integrating with more legacy systems arises, or as SaaS software is more widely adopted, the problem of ‘spaghetti integration’ emerges. Some businesses try to solve this problem with technology like enterprise service bus (ESB) or services oriented architecture (SOA), except that often results in ‘spaghetti in a box.’
Other organizations may recognize there are integration challenges, and opt to select SaaS software with published application programming interfaces (APIs) as the way to avoid data silos. The result of that decision is that they now have to manage integrations against those APIs, and access to them.
Knowledge about SaaS integration is a classic example of the stages of competence, and many organizations experience the stages of cloud competence. It’s after they have risen through stages one and two, and are at the stage three level of conscious competence that they have a solid understanding of cloud computing.
How can your business avoid spending too much time, cost and resources on the bottom stages of the cloud competence pyramid, and achieve maturity more quickly to ensure their SaaS strategy works?
How to make SaaS easier?
The first step is to develop a data strategy. This involves identifying what to do with a company’s data – how it will be accessed, shared and managed. There are a number of aspects to this, including:
- where to store master data and record business transactions,
- how to integrate the data between systems,
- who gets to use it and how.
The next step is to map out an integration plan. How are you going to integrate the platforms – cloud-to-cloud, cloud and on-premise – and what tools are you going to use. The integration solution you select should be able handle any platform integration issues – cloud, on-premise, mobile – as well as multiple integration points without creating ‘spaghetti’. The tool should also be able to act as a single, comprehensive API management centre.In the same way that a business manages its fixed assets, it needs to do the same with its data if it is going to survive in the information age.
One aspect of SaaS applications that gets overlooked is the software upgrade schedules. With on-premise software you control the timing of upgrades, with SaaS software that control is lost. As the number of SaaS solutions grows, so you have to develop an agile approach to dealing with a disparate upgrade schedule. That means changing your development methodology to support agile development techniques. Obviously the integration tool you use should support agile integration.
The saying that forewarned is forearmed can apply to companies that start moving their business software to the cloud. There are a new set of challenges that need to be addressed. There are tools available that can assist in tackling these challenges, but it also requires a new set of skills and approaches. Analysts are calling these news tools Integration Platform as a Service – iPaaS. In the heyday of enterprise software suites, businesses got used to the concept of hiring implementation or business process professionals. In the era of cloud computing there may be a new professional that organizations will need – the integration custodian.