Will you "rent" software in the future?

Once the preeminent belle of the high-tech investor's ball, companies that produce large client/server software applications for businesses have recently hit on hard times.

5 min read
What it all comes down to is a different set of values
to throw away or mobilize to use.

--Sun Volt, Questions

The answer to the above question is yes, but the answer is more complex than simply moving your application offsite.

Once the preeminent belle of the high-tech investor's ball, companies that produce ERP (Enterprise Resource Planning) software have recently hit on hard times. ERP systems are large client/server software applications used by companies to help organize and run their businesses. Until recently, many experts believed that the companies that sold these applications, such as SAP and PeopleSoft, were poised to become the next Microsoft or Cisco. With Fortune 500 customers virtually "running their business" on top of these applications, the strategic positioning appeared unshakable.

Recently, however, most of these ERP companies have disappointed investors with respect to revenue growth, earnings per share, and, as a result, stock price performance. Ask three people what happened to the ERP market, and you are likely to get three separate answers. At first, many pointed to the infamous Y2K issue, suggesting that CIOs were postponing ERP initiatives in order to devote time and resources to this more pressing issue. Yet as many companies mentally move past Y2K, this argument holds less weight. Others would suggest that the ERP market simply reached saturation. The idea is that more than 50 percent of the Fortune 500 have implemented ERP systems, and it is simply impossible to maintain unburdened growth.

Despite the potential validity of these previous two explanations, the high-tech community is much more infatuated with a third explanation: that the client/server software architecture has run its course. According to this thesis, client/server software has become overly burdensome. It is too expensive, it's too hard to implement, and it's too complex. Furthermore, with the Internet around, why not simply host these applications offsite as a service and let companies rent these applications and run them remotely?

Hence, the rise of the ASP (Application Service Provider) market. ASPs, such as US Internetworking and Corio, plan to run ERP systems in their own datacenters, reducing the support costs for the customer. The ASP also is the beneficiary of the coveted recurring revenue model, which is much more appealing than the previously sexy high-margin software license.

Should companies line up to rent ERP applications over the Web? And will the ASP market be the next hot industry to emerge from Silicon Valley? The answer is yes and no. You will access software functionality as a Web service. However, if you look at the opportunities created by the rise of the Internet, running SAP in a closet at your nearby telco falls short of what is possible. The architecture is troublesome, the economics are difficult, the feature set is misplaced, and the concept of "user" must be redefined. This is indeed a new paradigm, but it is much grander than most people are willing to admit. New Internet-based service providers will emerge, but they will look very different than their ancestor--the ERP software company.

For the sake of argument, take it as a given that running an application offsite makes sense in that it removes the installation and maintenance burden of the customer (ASP model). In this scenario, what is the optimal technical architecture for the Web-based service? In order to realize scale, one would likely argue that you would want one large "application" that is capable of servicing multiple customers simultaneously. Interestingly, most of today's ERP systems support only one company per application. This implies that for each customer, the ASP will have dedicated hardware and dedicated software--hardly the image of a scalable operation. Performance and responsiveness also will be a major issue, but today's ERP systems were designed for technical elegance, not raw speed. If you want to build a high-speed Web service, Amazon.com and Yahoo are your technical proxies, not SAP. ERP systems suffer from performance problems on a LAN. This will only be exacerbated on the Web.

ASP Economics
To make the ASP model an effective business, ASPs will need to charge a fee higher than the aggregate cost of the application, in addition to the cost of running, managing, and hosting that application. With one application per customer, ASPs will likely need to charge approximately 40 percent to 50 percent of the purchase price of the entire software application annually to have any hope of profitability. This may prove to be a hard sale, as the only way a CIO can justify such a play is to believe that once the application is outsourced, half of the IT group can be dismissed. However, CIOs, like most managers, are protective of their budgets. ASPs also may find themselves competing directly with salespeople from the software company they are reselling as a service (especially with respect to price). Even more competitive to the ASP model, ERP companies will one day wake up and find it strategically important to control the service themselves--an idea that Oracle already has voiced.

Features and purpose
ERP software primarily serves as a data repository and analysis tool used by a single company to help optimize the performance of its business. Software implemented as a Web-based service has a higher aspiration. Once the application is on the Internet, there is a huge incentive to connect multiple parties to the system to improve communication between companies. For instance, if you run financial software as a service, why not create an "accountant's window" whereby audits can be conducted directly from the Web. Additionally, HR services will connect to payroll systems as well as benefit providers. If the goal is to truly outsource, doesn't it make more sense to outsource the entire task through communication enablement rather than by moving a software application on a new premises?

The point here is subtle but extremely important. Web-based services are more about communication and multiple company workflow than the data-intensive functions of traditional ERP software. A few years back, analysts were very excited about the rise of intracompany workflow systems such as Lotus Notes. Inter-company Web services are 10 times more powerful. Companies that are connected together with partners, suppliers, customers, and service providers through industry-specific, Web-based services will be infinitely more competitive than companies that rely on internal systems for optimization. Why analyze and predict what your customers might order if you can be connected to that customer directly? Why spend tons of money on an HR system only to print a report and fax it to a service provider? Why not be connected?