The problem with design-build, as a general concept, is that the system is stacked in favor of the builder. Naturally, the builder does not see this as a problem. But the buyer should.
Let me back up a little.
According to Wikipedia, "Design-build is a construction project delivery system where, in contrast to 'design-bid-build' (or 'design-tender'), the design and construction aspects are contracted for with a single entity known as the design-builder or design-build contractor." Design-build is widely used for small- to mid-size construction projects because of the system's many advantages. Enhanced communication and a faster turnaround time are staples of the design-build system.
Design-build has a few drawbacks, as well. Major among them is the system's short-cut design process. And because the builder is also the designer, it is only natural to expect that the designs will gravitate to technologies most preferable (read profitable) to the builder. If you ask the stone mason to design and build your house, you can be certain that the house will have stone walls. Ask the carpenter - you'll get two-by-fours all over. You get the picture.
Design-build can be efficient if you are building a shed, or an addition to a brick house. But if you are a city government, looking to invest in a new stadium or concert hall, you will take a different approach. You will start by selecting an architect.
Why am I talking about construction in this article? Because everything I said about a construction system applies to any other system, including a GIS.
How does an organization go about initiating a GIS implementation? More often than not, it is by first selecting a vendor (i.e., a builder). This process is stacked in the vendors' favor, as discussed above.
In my 18+ years of GIS implementation experience in the New Jersey region, I have witnessed hundreds of GIS installations, mostly in local government. Almost all of them were procured and developed through the design-build process, as if no other project-delivery system existed. Buyers would issue a home-grown RFP, collect apple-and-orange responses and then select a design-builder (usually a software vendor or an engineer). With increasing frequency, the buyer would throw out all responses because none of the respondents could correctly guess what the buyer was looking for.
If a contractor selection is made and the project does move forward, the implementation jumps straight into data development, because the vendor selection has automatically selected the platform. What happens next?
Project delays and cost overruns are common. Many systems do not meet the buyers' current and future needs. Many are shelved. Many are never maintained. Many meet with resistance from their intended users. Why is that? Because of a poor, rushed or entirely missing system design phase.
There is a better way. Begin the process by selecting and hiring a GIS architect. GIS architects know the industry inside and out, follow technology trends and know the vendors' strengths and weaknesses. An independent GIS architect will study the owner's needs, help develop a program, prepare an RFP, select a vendor and oversee the system implementation.
It is high time to recognize GIS architecture as its own specialty and GIS architects as a crucial and integral component of the GIS design, development and implementation process. Nobody disputes the need for an architect in building construction. Nobody disputes the architect's fees, either (running between 6.5% and 30% of the total construction cost). This is just how it works. And it does work. Just look around - there are so many beautiful buildings. Even the ugly ones function as intended. Rarely does a building collapse on its own.
Not so in GIS. When it comes to GIS, all too often organizations do the equivalent of hiring the bricklayer to do the design, all too often with devastating results. The GIS landscape is littered with Frankensteins, a wasteland of wasted effort, where bitter early adopters often claim that GIS stands for "Gee, I'm Sorry."
If you are planning to spend $1M on a GIS, budget an extra $65,000 (at a minimum) for a GIS architect. You won't be sorry.