Companies are taking a strategic view of integration challenges to ensure good results.
by Carol Hildebrand, May 2009
Changing business conditions are likely to bring on more—not fewer—integration projects, and companies need to think strategically about their goals and areas of differentiation before the work starts. R. “Ray” Wang, vice president of enterprise applications strategy and contract software negotiations at Forrester Research and author of the blog A Software Insider’s Point of View talked to Profit about reducing ERP risk and the challenges of application integration.
Profit: We are hearing more about incremental enterprise application installations, rather than the “big bang” approach. What’s the value of the incremental approach, and are you seeing more of that type?
Wang: I think we started seeing more of it in the last two quarters, as companies grew more risk averse about committing resources to large projects and implementations. The key things about incremental implementations are really focused deliveries, a lower price point, and minimal disruption to businesses. On top of that, many companies already have large enterprise applications, and their projects naturally tend more towards adding incremental functionality. I’m not saying that big bang projects don’t work, but I don’t see them happening regularly unless there’s a major regulatory upgrade.
Profit: Are you hearing about more companies having success with vanilla implementations?
Wang: Yes and no. Vanilla works great if you are trying to catch up or improve on applications such as finances or CRM—processes that are essentially commodities for many companies. The trick is knowing which processes are commodities to your company, and which will add value if customized. Does it make sense to have the most unique HR process? Probably not. But say you have an innovative way of handling multichannel order management—that’s where you want to do something special through customization.
Profit: Will application integration remain on the front burner as a cost reduction/business value project?
Wang: I think companies tend to pursue integration for three reasons. First, they want to leverage their existing investment—they don’t want to start over, but they do want some interactivity between their applications. The next reason is consistency of metadata and data processes. This is necessary if they want to share information across applications. The third reason is to get more value, to harmonize applications so that they can work as a virtual suite. If you have best-of-breed applications that need to talk, integration will continue to be on the front burner. But more than cost reduction, it’s about being able to be strategic, to use information to get an edge.
Profit: With today’s economic volatility, how can companies minimize their risk profile in these areas?
Wang: It’s all about good design principles—you must have good upfront planning. Instead of integrating applications on a case-by-case basis, you need to have a road map for applications all the way to projected decommissioning—you don’t want to integrate stuff that you are getting rid of, but you do want to account for processes and the functionalities they support. You need to think of the overall data model and how to handle design services. Think about what roles users will play, what processes play a part in which data requirements, and what outcome you are trying to achieve. If you are trying to get information to frontline employees, start there and reverse-engineer back to the datasources. It’s like an information supply chain, and you need to figure out the best way to get the goods to the right people. Another way is to see what others have done. Build on their experience and eliminate problems ahead of time by examining lessons learned and best practices within your industry.
Profit: Having talked to many companies, what are some of the differences you see in how companies approach integration?
Wang: The first step that many companies take is to figure out whether they want to use integration packages or go the custom code route. If you work for a company that does a lot of custom development and has experience writing code, that type of company will tend to want to tackle integration itself. If a company is used to packaged applications, chances are it’ll want as many templates and processes and models as possible. It does vary by industry—telcos, financial companies, and insurance firms generally build for themselves because they are always looking for an edge. Retail and public sector organizations, on the other hand, tend to gravitate toward packaged applications.
Profit: How do you see application integration evolving over the next five years? For example, does the Oracle strategy of offering preintegrated software packs make sense?
Wang: Integration is now a high-value act—it seems that the #1 priority we see on every Forrester survey is that companies want their applications to talk to each other. Once that happens, companies can take advantage of it and innovate. With that in mind, there’s been a rush by vendors to build integration pieces. I think we’ll end up seeing a lot of people build the same set of integration connectors, and there will be fallout as those pieces move down the technology curve into commodity status. As for the PIPs [process integration packs], I think they are an interesting approach if you’re looking for that level of integration. If you are an Oracle customer by choice or by acquisition, you can reduce costs by using the packs—essentially, letting Oracle handle the integration for you.