Software as a Service (SaaS) is cheaper, quicker and easier than traditional on premise applications. Right? Well, certainly according to the advocates of the on-demand computing movement it is. You can also switch a SaaS application off if you do not like it. You are not committed to a multi-year implementation process that you cannot back out of, so what is to stop any organisation just diving in and having a shot at SaaS?

The theoretical answer is nothing, although the real-world answer might carry more caveats. Forrester Research carried out a study of customer relationship management (CRM) deployments, which included SaaS roll-outs, for its report "Best practices: the smart way to implement CRM". What the analyst firm found was that there is a need for planning and piloting with SaaS just as there is with traditional application roll-outs. "With more frequent upgrades, faster deployment, lower upfront costs and high acceptance by employee end-users, customer demand for CRM SaaS applications shows no signs of slowing down," concluded Forrester analysts William Band and Peter Marston, but added, "Successful implementation and integration requires that you follow sound practice."

These are fairly standard practices, ones that can be carried over from the world of on-premise software. Why are you doing the implementation? Do you have a business case in place? Do you know what you want to get out of the deployment? How will you measure success or failure? When considering cost, have you taken into account more than just the up-front monthly rate? What about software licence fees, internal labour implementation costs, professional service fees, user training expenses, mobile and offline/online system access, industry-specific functionality, storage capacity fees and premium help desk support? Do you have a tough enough set of Service Level Agreement (SLAs) in place?

One company that has considerable experience of SaaS deployment is Textron, a manufacturer of helicopters, aircraft, fastening systems, tools and components, and a provider of financing tools, which has adopted SuccessFactors as the company standard for on-demand employee performance management. Textron embarked on its deployment as far back as 2001, making it a veteran among SaaS users of any supplier in any business category.

While the number of enterprises deploying SaaS is increasing, seven years ago they were few and far between. "Perhaps it was brave of us," says Will Roth, director of organisational development at Textron. "SuccessFactors did not have large customers at the time and certainly not any that you would have put in the same league of complexity as us. We were rolling it out across a dozen business units. We had the choice of going to a boutique shop who could come up with something that would be close to what we wanted up-front, but then you would have the issue of how you maintain that going forward. If you go to the leadership of the company and tell them how much this thing is going to cost if you run it internally then you are lying to them because you do not really know."

That said, cost and ROI were factors in making the move to SaaS for Textron. While some analysts have questioned whether SaaS return on investment (ROI) is viable over a certain period of time, usually pinned at around five years, Textron's seven year SaaS engagement does not bear this out. "We saw ROI in the first year," says Roth. "You have to think about the number of people we would have had to put on a project like this ourselves. Hardware itself is not the cost the problem is the people you have to have people out there maintaining your system. That is where it gets so intricate. Every time you want to add a new feature, you need a lot of people. You cannot easily put a definitive number on what it costs to do this sort of stuff internally. You do not have to be an actuary to realise that there are human resources costs beyond that of the person working for you. Some of that stuff lives longer than the application itself."

There are of course now many more SaaS providers than there were seven years ago when Textron began its investment. Roth advises that anyone thinking of taking the plunge should consider a couple of factors to eliminate unsuitable suppliers. "A supplier can have a product that is not quite as good as others, but two things will make it suitable," says Roth. "Security was one thing that we looked at. If you are doing any of this stuff you really do have to look at security. It is one of the most important filters when choosing a supplier. Three quarters of the possible suppliers will not pass examination of security credentials. There are a lot of suppliers who have internet applications, but when you do a bit of digging then they do not have the right level of security. Customer support is another factor that differentiates suppliers. SuccessFactor's CEO was fanatical about customer responsiveness and support. He created a culture in his company about this. I was almost scared that when SuccessFactors got to a certain size then they would not have that anymore, but he has managed to keep it in place."

Textron chose not to go through an extensive piloting period for the SaaS roll-out. "We actually went big bang," says Roth. "We did do a very small pilot with a very small group of people. I am not sure I would even call it a proper pilot. The idea was not to stagger the roll-out. We looked at 50 people using the system. It is easier to get the feedback you need by using real world users than by using test users. We then went with the whole enchilada. If we'd rolled it out one business unit at a time then everyone would have told me that they were somehow different from everyone else and needed something special. We are still rolling out features, so we are not that much different to new customers in that respect. Again we are doing a small pilot, but come December we are to go with the whole thing again. If you plan it slow, it is going to go slow. The best return on SaaS is going to come from having everyone on it."