As in all things, timing is everything. If you move too soon, you're so far ahead of the market that you won't recoup your investment in a timely manner. If you move too late, then by the time you do move, the market opportunity has already passed you by.
But if you ride the curve of adoption for new products and business models just right, profitable revenue streams are sure to follow.
For example, while the managed services trend has been around since the early days of the dot-com bubble, it's only now that this model is beginning to find mainstream adoption among customers. Software as a service is on a similar trajectory in that there are some high-profile success stories, most notably in the CRM space, but by and large the rest of the IT application space is still playing with the concept.
The questions facing solution providers are: When will the market tip more decidedly toward the software-as-a-service model and how should solution providers think about responding to this eventuality?
Now to be fair, despite the hype, software as a service is not going to completely displace on-premise software deployed on the customer's servers. But it is going to take a larger percentage of the marketplace, probably to the point at which 50 percent of all applications will be deployed as a service.
Given that software as a service today probably accounts for less than 10 percent of all applications sold, it's pretty clear that we're about to reach some sort of tipping point in terms of the adoption of software-as-a-service delivery models.
That tipping point, driven largely by customers looking for less expensive ways to try out software without having to tie up huge sums of capital, has huge implications for solution providers because it not only changes the dynamics of software licensing, it usually leads to fewer hardware sales and services revenue that is derived from professional business consulting rather than technical implementation services.
So how do we know we're at that tipping point? Well, there were two pretty compelling pieces of evidence in the news recently. The first came in the form of news from Microsoft that the company has finally figured out how to deliver a Web services layer that isolates data models from the underlying application infrastructure.
This will allow Microsoft and others to more easily deliver software upgrades as a service while keeping the data schemas of those applications intact. This layer puts some teeth in Microsoft's ambitious Live plans to deliver hosted applications directly to customers and in combination with its partners (see http://www.microsoft-watch.com/article2/0,2180,1943281,00.asp)
Microsoft 'CRM Live' and 'ERP Live' may not take the hosted-application shape that many company watchers, partners and customers have expected. Click here to read more.
But if Bill Gates' support for the concept isn't enough to convince you, then maybe you'll be convinced once you consider that the second richest man in software agrees with the richest guy in software. Oracle's Larry Ellison has begun telling Wall Street he expects the Oracle business model to shift to a subscription model as part of the general shift to software as a service.
For solution providers, this means the time is now to decide how your company is going to participate in this shift, which comes down to three basic options. Door number one turns your company into basically an agent representative that gets a nominal fee for signing people up to use a service and hopefully getting compensated if they renew. The former model is represented by Salesforce.com, while the latter model is more often associated with NetSuite.
Door number two might be a little more compelling from a profit point of view, but it requires a greater amount of capital investment. In this model, you basically concur with Microsoft in that elements of any application are going to consist of pieces that reside on clients that communicate with servers running applications that are run either by Microsoft or its partners.
The third option is to bet that client code is going to be pretty much irrelevant in the future and, unless you have original intellectual property that fundamentally differentiates your application in the market, you'll pretty much be run over because profit margins on generic services will be pretty thin.
There's nothing that precludes using this approach on either Linux/Unix platforms or Microsoft platforms, but the Microsoft platform is decidedly more client-centric. That approach comes with a number of philosophical questions about the value of a significant amount of code on the client, versus a more server-centric approach that theoretically is easier to deploy and maintain.
Whatever path to the future you decide to pursue, the most important issue is balancing the amount of capital investment required to be successful against the profit potential of any given approach. You could just as easily find yourself heavily invested in a high-margin open-source application that has limited revenue opportunities as you could supporting a set of high-volume applications that produce little profit because there are too many providers.
But whichever way a solution provider decides to jump, one thing is for certain: Solution providers are running out of time to make the decision because unless they start to adjust their business models and technology platforms today, the market will surely pass them by this time next year. Every model has its upside and pitfalls; the only question is not whether to swim, but rather which shark you should ultimately swim with.
Michael Vizard is editorial director of Ziff Davis Media's Enterprise Technology group. He can be reached at michael_vizard@ziffdavis.com.
Commentary 
