Tribal Knowledge - Enterprise: How long is a piece of string?

"New technology renders existing products, services and training obsolete on a daily basis. Some will tell you that this wipes the slate clean - well, think again. I'm here to tell you that all that you know - you need to remember. This is a monthly column dedicated to dispelling hi tech myths and sharing life's lessons."

This month: Enterprise: How long is a piece of string? (The first in a series)

Like a the buzzing of a summer insect, there is a never-ending tug of war between vendors, analysts and IT professionals to define exactly what constitutes an enterprise solution (enterprise class, enterprise grade, enterprise ready, etc.). The following mini-series of columns will explore various aspects of the enterprise, have a little fun at vendors' expense and maybe point out a few perennial truths along the way to help readers make better decisions (and avoid getting a world class enterprise bug up their ASP).

Despite the many views on the subject most would agree on two points; enterprises are big and they're where the money is. So, it must follow that the bigger the software solution the better - right? Well, sometimes yes - but let's not get too far ahead of ourselves - is bigger always better?

This excerpt from an early Saturday Night Live skit spoofs both the market's emotional desire for simplicity and the absurd lengths suppliers will go to feed this desire. Humor aside, history is littered with examples of companies whose hubris led them on a quest to take all of their customers' needs (and thus their money) and put them into one product. These quixotic quests have often yielded nothing more than monolithic offerings with inefficient cost structures and poor quality.

Whether it's the authoring tool vendor that wants to move into content management or the DBMS vendor that wants to replace the file system or an automobile manufacturer that has had own dealerships, factories, steel mills and even the mineral rights for the raw materials, suppliers can be counted on to go (at least) one step too far in pushing their offerings.

The illusion of economy of scale

Most are familiar with the basic notion of economy of scale, e.g. the greater the volume or level of activity, the more efficient pricing, production and distribution should become. The underlying assumption is that there is one fairly homogeneous set of tasks/products/services that are being scaled. Economy of scale does not work if this assumption is not met.

A subtle twist to this concept emerges whenever there is "convergence". One way vendors attempt to grow their target markets is to merge distinct categories and call them one.

Content management is a prime example of where convergence DOES work and when it DOES NOT.

The following chart drills down on specific aspects of content management where the ultimate value of unification is questionable at best. Again, in the interest of those readers that feel strongly that the one stop enterprise content management system is the end all - this table highlights those areas that DO NOT scale - this is not a comprehensive list of all aspects of content management.

Figure 1 provides and interesting view into the differences (and the similarities) across content management categories.

Best of breed approaches where technology is acquired a la carte comes with some obvious risks. Assuring that all of the moving parts work well together and maintaining that integration over time has historically been a significant challenge. Occasionally, the mere fact that a single company has stepped up to own and manage the integration of multiple moving parts - even if there is no economy of scale - is enough to justify abandoning the best of breed approach. This is true even if the unified offering's only redeeming quality is that there is one company to call when something goes wrong. However, there are a number of factors to consider when weighing one approach versus the other.

The following trends are strong indicators that an "all-in-one" approach is NOT likely to outweigh a best of breed approach.

• Requirements are evolving for one of more components. Development environments where boundaries are not neatly defined between components lead to turf wars within single provider organizations.
• Software categories are evolving for components. In addition to turf wars, partnerships, requirement collection and overall strategy are likely to collide slowing market responsiveness.
• ROI for integration is still being refined and improved. Until the ROI for integration is well understood, flexibility and the option of dropping or swapping components is essential to maximize technology investments.
• Potential for development and resource conflict. Without a single vision and a shared sense of resource allocation, the trap of keeping too many projects going for too long drains resources resulting in slower release cycles and lower quality.
• All-In-One solution has no business track record or is not profitable. If an All-In-One vendor does not have a track record of making money off of an integrated solution, then it is likely that one or more of the components will be retired. This results in the worst of all cases - mediocre components with no out of the box integration.
• Best of Breed Vendors have existing integration with ongoing commitments to customers. If best of breed vendors have already integrated their products and have committed to supporting this integration going forward for other customers, the likelihood that the integration is complete and will remain up to date is high.

Is an all-in-one approach ever best? When mapping products such as Microsoft Office to the criteria listed above, one can see why the integration is very appealing. There are clear boundaries between the products and well understood ROI and use cases on how the components should inter-operate.

End-to-end is code word for computing silo. There are arguments to be made on how much functionality is too much and when is the best time for a consumer organization to trade off between single point solutions and best of breed. However, there can be no doubt that every vendor must integrate with other commercial software vendors.

There is no vendor that offers authoring, management, layout, formatting and delivery across all channels and media. An organization that has put its muscle behind an all-in-one or end-to-end solution will not have the same investment in integrating with your other suppliers as a vendor that has third party integration as a core part of its business model and product architecture.

So, to draw this installment to a close - vendors that rollup well understood applications are likely to offer something of real value - bigger is better. However, like a loaf of bread, end to end solutions can rarely be un-baked. If there is no consensus on best practices across large deployments - best of breed should provide maximum value and flexibility. So, just as I repeat to my one year old daughter, Adriana "convergence driven by technology - all-in-one bad. convergence is driven by application fluency of install-base - all-in-one good."

Happy Independence Day To All!!!