Enterprise-wide integration. Who among us has not worshipped at this altar? The vision of all IT systems communicating seamlessly has undeniable appeal. So it’s no wonder enterprise-wide integration is part of the e-business vision: with Internet technologies as the glue, we will finally link all systems together. Ah, what a glorious day that will be!
The only problem: the vision is a chimera, a false goal that encourages, at best, frustration over dashed hopes and, at worst, misallocation of IT resources.
I have long known that enterprise-wide integration had reached motherhood-and-apple-pie status, but I didn’t realize how much it was influencing e-business thinking until recently. My company’s latest industry survey, Survey of e-Business and IT Practices, showed that much of our industry suffers from the delusion that total integration is achievable.
Tellingly, the results also put the lie to enterprise-wide goal by showing that the real value of integration lies in limited connections among complementary systems.
The appeal of enterprise-wide integration is easy to see: Instead of relying on isolated application systems running independently (as “islands of automation”), businesses would benefit from a network of linked, interdependent systems. Where output from isolated databases can be assembled into coherent reports only with great effort, integrated systems would produce consolidated data as a routine product. Other benefits would include elimination of redundancy, optimization of software licenses, and reduction in infrastructure costs.
That’s the theory. The reality is that enterprise-wide integration is too costly and too complex to be a legitimate goal for most companies. The real goal should be to integrate selected systems only when benefits outweigh costs, which generally happens only when the systems are somehow complementary.
This theory-practice dichotomy shows up in the survey, which targeted medium to large companies in a variety of industries. When respondents were asked to rank the leading impediments to e-business in general, they ranked integration with legacy systems as the number-one issue. Meanwhile, almost three-quarters of the same respondents say their core systems are not enabled for e-business.
In other words, respondents believe they should be integrating e-business systems with other systems (the motherhood-and-apple-pie argument), but they’re not doing it (the practical realities).
What they are doing is integrating applications on a selective basis.
Among those respondents selling products over the Web, 54 percent have integrated their customer-service application with their e-commerce system; 38 percent have integrated sales and e-commerce applications; and 29 percent have integrated e-commerce with their financial apps. Linking these systems makes sense, since they share a lot of common data.
Respondents are much less likely to have integrated e-commerce systems with applications like procurement, manufacturing, and field service. This makes sense, too, because these systems don’t share much data at all. (One-third of respondents are using e-commerce as a stand-alone app with no integration.)
In reality, then, e-commerce integration remains limited, certainly far short of enterprise-wide. If e-commerce were integrated throughout a business, manufacturing could feed orders directly into production and materials planning, with logistics scheduling shipping, procurement ordering products and raw materials, field service providing customer histories, and so on. Respondents recognize the limitation, ranking “lack or difficulty of integration with other applications” as the number-one cost of doing e-commerce.
Does this mean e-commerce is a failure that respondents are ready to abandon? Hardly. First, almost half remain “satisfied” or “very satisfied” with their e-commerce efforts. Second, 92 percent plan to increase their e-commerce practices, with the remainder planning to maintain their current level of use. Online selling’s benefits, even with limited or no integration, still justify greater investment.
Other online applications show parallel results. E-procurement, for example, has been integrated with financial applications by more than half of those buying supplies online. (The next most commonly integrated applications are logistics and manufacturing.) No integration remains common — 47 percent of those using e-procurement have not integrated it with anything — and, as with e-commerce, respondents ranked lack or difficulty of integration with other apps as the number-one cost of buying supplies online. But, again following e-commerce, respondents are satisfied and plan on doing more of it in the future.
The pattern continues with other e-business applications like e-CRM (electronic customer-relationship management) and e-SCM (electronic supply-chain management). E-business applications are integrated with a select few complementary applications (and even those not universally). Application users bemoan the cost and difficulty of integration but are sufficiently satisfied to plan on using the apps even more extensively in the future.
Clearly, enterprise-wide integration is a theoretical goal that crumbles in the face of business reality. System integration, enterprise-wide or otherwise, must pass muster before the cost-benefit tribunal. If it costs more than it returns, or if other investments will return more, it fails the test and should be foregone. No use wringing our hands over this — it’s the way business is done. In fact, we should be glad such a simple principle can guide our decisions.
Mythology students know there’s only one way to deal with a chimera: Cut off its head! Leave enterprise-wide integration to the dreamers and the pundits. Focus solely on the integration that makes sense for your company, and don’t give a second thought to the integration that doesn’t.
Chris Pickering is president of Systems Development Inc., an IT research and consulting firm. He also is a senior consultant for the Cutter Consortium. His latest industry study, the Survey of E-Business and IT Practices, may be ordered now from Cutter at www.cutter.com/consortium.
Editor’s note: This story originally appeared on Datamation, an internet.com site.