Evaluating the success of IT implementation in organizations

One of the recurrent threads in this both week’s material and last week’s has been the relative success organizations can achieve with implementing new elements into their supply chain.  Whether the changes are procedural or technological, they are fraught with their own respective issues that can accompany any kind of disruption to business as usual.  With lean and implementing the Toyota production system, some gains can be had from only partial (admittedly far from ideal) implementations.   However, changing the IT backbone of an organization is a much more difficult move to make.   Work does not just become less efficient, it can slow down or stop entirely if legacy systems are unable to communicate with an ERP or the SOA tools used to link those legacy systems are incompatible.

So the question becomes, under what circumstances are these upgrades necessary, and how can they be most efficiently executed?  I’ll first stipulate that the usual change management issues are likely to arise, just out of the novelty of changing a process.  Beyond the more people-oriented way of thinking about things, though, is the need for the IT and business sides of an organization to come to a mutual understanding about the feasibility and limitations of tacking on monolithic new systems like an ERP.  Despite the best intentions and efforts, up to 75% of ERP implementations fail[1].  The way the issues are described in an article by Cynthia Rettig, it looks to be an issue of complexity and scaling.  If monumental organizations are to connect their many disparate systems, the degree of interoperability required grows both with the demands of the system as well as the number of factors that have to be integrated.  So, logically it makes sense that SOAs succeeded the massive ERPs in some fashion.  The tailoring required to make this successful though makes it cost prohibitive.  Each company needs its own specific set of SOA modules. This brings us to cloud computing as a solution. 

Cloud computing begins to solve some, not all, of the issues of complexity.  Rather than require all of the various parts within an organizations local network to be able to communicate, cloud computing takes advantage of existing infrastructure, and with various service rental models allows maintenance and administrative costs to be deferred.  As companies integrate the relatively simpler to implement cloud solutions into their SCM IT systems, I would expect adoption rates to increase.

This brings up an issue and a tradeoff in my mind, however.  What is the difference is realized competitive advantage of fully functioning ERP versus some of the newer SOA and cloud solutions?  The latter two come off as being more likely to succeed due to the ways in which they defer some of the issues of complexity.  Especially with cloud solutions though, they are much more accessible to many organizations which would dilute some of the gains that could be had from adopting a cloud IT solution to replace older, legacy systems.  Though ERPs have developed a stigma for being hugely risky investments, it stands to reason that their relative uniqueness could retain some of the competitive advantage that is lost to more generally, accessible cloud solutions.

---

[1] http://sloanreview.mit.edu/article/the-trouble-with-enterprise-software/