Intelligent Enterprise

Better Insight for Business Decisions

RSS
Webcasts
Whitepapers
Subscribe
Home

April 22, 2003

The Future of Enterprise Applications

An Intelligent Enterprise Roundtable

by Justin Kestelyn

Continued from Page 1

Graf: SAP has been involved in outsourcing for a long time. In our experience, customer acceptance depends greatly on the business process involved. If you're outsourcing things that customers don't see as being strategic — payroll runs, sourcing of nonproduction goods, and things like that — it works well. Now, we're even looking at parts of processes to outsource — the content part of the sourcing equation, for example.

Based on that observation, outsourcing is not bound to a specific market segment — actually, it works across all segments.

Gassner: It's pretty simple: Some people want to rent and others want to buy. You have to support both models; some customers would want to implement their apps in a hosted, outsourced way, and other want to have them on site.

I agree with Cliff that configurability is very important. Yes, customers in a hosted model will live with less customization, but they still need heavy configurability because even in the hosted model, one size doesn't fit all.

Hensarling: The driver here is that most companies are looking for lower total cost of ownership, and there are some "in-between" ways to get it. We're seeing customers respond to that need by bundling services with deals so they can get more verticalization at a known cost, and by taking on maintenance themselves. Whether the environment is hosted or not, that need for lower initial and ongoing costs is the key.

IE: Most observers agree that the midmarket offers the greatest growth opportunities for enterprise application vendors. How will customer requirements in this segment influence the way your technology is developed, packaged, marketed, and implemented and maintained?

Graf: Small and midmarket customers have already had tremendous influence in our solution development and offering. Today, we give our customers specific solutions created for organizations of specific sizes. Most important, however, is a strategy that allows companies to continue to grow — and the solution grows with them. In addition to ease-of-deployment and ease-of-use, these are the key requirements, especially for smaller customers.

Hensarling: Having lived in the midmarket space for our whole 25-year history, we're very attuned to that challenge. We look at it as a buying criteria that works in the midmarket, but also works in the divisional play for certain larger customers. They can get a clear sense of what it's going to cost them vs. the return.

Gassner: I echo those thoughts for the midmarket, but maybe not so much in the small market. Full functionality is often needed in the midmarket, so the trick is to make things configurable, go in quick, get the basics up and running, but still have the total functionality in there so they can turn it on as they go along.

IE: Let's revisit Lenley's point about achieving granular ROI. PeopleSoft, SAP, and J.D. Edwards have announced strategies for creating open integration frameworks in which application components will be orchestrated as defined business processes occur. Oracle, on the other hand, adheres to the monolithic architecture view. Which alternative will hold sway in the next few years? And what role will Web services play in each one?

Godwin: The characterization of Oracle following a "monolithic" approach is a misperception, which may have resulted from the fact that we strongly advocate the benefit of an integrated suite over integration-intensive, best-of-breed strategies. We may have created the impression that we're not paying any attention to integration, but actually, we are. It just hasn't been the core of our public message.

We're supporting XML-based message integration with trading partners and continue to support emerging integration standards, so I don't think Oracle represents a pole of opposition to the idea of standards-based, Web services-oriented integration. There are old-fashioned types of integration that are still relevant in many cases, but there's also a lot of leverage in being able to support new open standards for integration and business process execution.

Gassner: In our view, monolithic architecture never really held sway, especially in larger companies. People implement in bite-sized chunks; they might have a CRM project one year and a supply chain project the next, so they're going to implement and upgrade gradually. That's why PeopleSoft apps have always been architected as components that you can fit together, implement, and integrate.

I agree with Cliff that Web services aren't the only answer, but we see it as the most important one going forward because a standards-based approach is going to make everything easier. It's an oxymoron to try to differentiate on integration; you want to be a standard on integration so that everybody can integrate with you.

Web services is going to grease the skids in a lot of places, but we definitely feel that componentized architecture — and our AppConnect product facilitates the process of building these composite applications — has a big value because it meets the way large companies do business. It's also workable for the midmarket; we just have to preintegrate our components and ship them in a box.