I get asked about the pros and cons of application suites versus best-of-breed applications at least once a week. Vendors, IT professionals, analysts and business users have been debating this question since the first suites appeared in the late 1970s, but somehow it remains unresolved. The two approaches tend to alternate in popularity, with the cycles influenced by some mysterious combination of the economy, management philosophy, technology trends and vendor marketing campaigns. The best-of-breed strategy was last in the ascendancy in the late 1990s, but then the collapse of the internet bubble and the subsequent recession seemed to shift prevailing IT sentiment toward centralized control and single-vendor suites. With the global economy recovering, companies are once again shopping for software and arguing about application strategy.

The irony is that, while the participants in this discussion very often have a fundamentalist viewpoint, almost all organizations have a hybrid reality. Regardless of what the stated application strategy is, I can hardly think of a company that doesn't have a mix of suite-based and niche applications. Almost no one has the ability or desire to replace an entire application portfolio, so even in the most comprehensive suite deployment, there are always some legacy and specialized applications left in place. And then, over time, applications just seem to creep into the portfolio because of acquisitions, maverick purchases and tactical projects.

I typically get called in to referee these application strategy fights when companies are embarking on a new selection or major upgrade. They're concerned about the long-term implications of their decision and struggling to make sense of the competing vendor claims. For years, I have done the "usual consultant thing": I assure both camps that there's merit in their arguments and that the right strategy depends on their industry, company culture and maturity. I then take the selection team through an exercise to explore the advantages and disadvantages of suites versus best-of-breed applications in their environment. The process goes something like this.

First, we discuss the suites where the big selling point for users is generally the process integration, and the possibility of finally getting to a consistent data model and a common user interface. The IT organization tends to like the idea of dealing with one major application provider and having a consistent architecture. Senior management, being risk averse, is likely to favor the large, well-established suite vendors, and they love the story about "built-in industry best practices."

Then we talk about the downside of suites - the time, cost, risk and disruption of a multi-year implementation project - and the fact that all that hardwired integration often comes at the cost of increased complexity and rigidity. Integrated suites also require a level of "readiness" that many organizations simply can't achieve. Unless there's a broad, internal consensus and a willingness to replace existing processes with those supported in the software package, the suite approach doesn't make much sense. And then, finally, there's the question of functional fit. A suite is only a reasonable option when it has functionality that meets the requirements of the business and industry.

Having thoroughly confused the team about suites, we then move on to examine the pros and cons of the best-of-breed approach. The appeal here is functionality and flexibility. Specialist application vendors have the luxury of focusing on a single business process area or vertical industry - oftentimes, they focus on both. They can typically offer more comprehensive functionality and a user experience that is more targeted and appealing than the general-purpose suite. For the individual department, the benefits are obvious: It gets a vendor that understands its needs, the functional fit is better, and the implementation is faster and easier, with a much greater likelihood of user adoption. From IT's perspective, these are much easier projects because the scope is smaller, the cost lower and the time to value often much shorter. The architectural vision behind this strategy is one of a more loosely coupled portfolio of applications that sacrifice some level of integration in exchange for a more tailored, functional fit, and the ability to more easily change individual applications to fit the needs and priorities of the business.

Unfortunately, the best-of-breed strategy also has plenty of disadvantages. The big one is that enterprise business applications simply don't "plug and play." Despite the claims of middleware vendors and standards advocates, integration is getting harder, not easier. The applications themselves are more complex, and user expectations about integration have changed. Users are no longer satisfied with the ability to pass data between applications. They now demand real-time process and data integration - even a common look and feel. That's hard for the largest application vendors to achieve within their own products, and nearly impossible for the typically understaffed IT department to pull off.

Another significant problem with best of breed is that the specialist vendors are often much smaller than their suite counterparts. This may raise concerns about the company's financial viability, and it frequently means that there are limitations in available resources and geographic coverage and support.

At this point in the discussion, I generally find that everyone is unhappy because, regardless of the strategy they prefer, they want some of the characteristics of the others. It turns out that best of breed or integrated suite probably isn't a very good basis to guide application selection.

Gartner subscribers and regular readers of this blog have heard me and Dennis Gaughan refer to an initiative called "pace-layered application strategy." This is an effort by a number of analysts to develop a new model for evaluating and managing business applications. We encourage companies to take a fine-grained view of their applications and use a set of characteristics to assign these apps to one of three levels: systems of record, systems of differentiation or systems of innovation. The purpose of this is to help companies do a better job supplying appropriate applications to support the business objectives, and shift more of the available resources toward growth and innovation.

One of the early outcomes of this work is that it appears to help organizations have a much more rational view of applications. Rather than striving for some monolithic strategy of "suite first" or "every department for itself," it advocates looking at each application and considering the nature of the business process it supports, any audit or regulatory concerns and the probable rate of change. The result may suggest a packaged application from a large suite vendor, a niche application from a specialty vendor or a custom-built composite app. The point is that most companies' application portfolios will, and should, contain a mix of these application types. We would like to help them get there by design, rather than by accident, equipped with a set of governance processes and "connective technologies" that allows them to maximize the value and minimize the chaos.

Source: Gartner