One of the most difficult questions that consultants are asked when trying to start a Business Intelligence project is, “What do you estimate the cost of this effort will be?” It is perfectly understandable that any executive pondering whether or not to embark on this journey would want to know how big of an investment would be required. On the other hand, the consultant’s classic answer of “That depends,” is also completely understandable. Without learning a lot more about the number of source systems, the quality of the data, the desired architecture, the amount of effort to be contributed by the client, the availability of technical documentation and/or people who can answer questions about the data structures, and a host of other very significant factors, it is unrealistic to expect a consultant to provide any sort of reasonable estimate of what the cost of the vaguely defined endeavor might be. It seems that many executives are reluctant to make a significant investment in an assessment phase, but consultants are usually unwilling to give away a large block of time of their valuable consultants without being compensated for it. How are the two parties to get to a point where they can move forward with any confidence or comfort? In this article, I’ll describe an approach that has worked well in a variety of situations.

One of the difficulties in coming to a reasonable shared understanding is that people on both sides of the chasm usually lack a significant understanding of the important issues and decisions that those on the other side need to address. For example, the executives and business users may not have a good understanding of important concepts like dimensional modeling and the strengths and weaknesses of various data warehouse architectures. On the other hand, the Business Intelligence consultants may understand the pros and cons of various technical options, but unless they have deep experience with a specific company, they may not appreciate the specific needs of the various business users and executives. Without that strong understanding of the business priorities, consultants are likely to take a “put it in the data warehouse just in case someone needs it” approach, which unnecessarily drives up the costs and reduces the ease of use of the resulting systems.

In order to come to a sensible agreement, the consultants and clients should begin with a high-level discussion that identifies some major options and the rationale for those options. But this will still fall far short of answering the question of what should be delivered and a reasonable estimate of how much it will end up costing. That is where the agreement on a process comes in. To understand the process, each side should think of building a road through a forest. The first road that gets built is just a roughed-out dirt road—and it is important that everyone understands that this is not intended to be a long-term solution. It should deliver some value, but more importantly, it allows everyone involved to get a much deeper appreciation of the challenges and benefits that will be involved.