For example, consider one of the most common requirements of a decision support, marketing analysis, or data warehouse environment: Be able to answer any reasonable business question with no worse than overnight turnaround. How do you translate this into a tangible deliverable? The point here is you might not know exactly where you are going, but some quick introspection can save you a lot of trouble down the road. Below are some simple questions to ask yourself before checking references.

You can download excellent powerpoint slides on Marketing Strategy and Marketing Management HERE.

Question 1: How Is Your Risk Tolerance?
This question has many facets. There is tangible business value to building a info-structure or implementing "relationship technologies." This implies there is risk in not being successful, or in only being partially successful. What is the business impact of a delayed implementation? If the product you choose is difficult to set up and tune, delays are likely. If you are relying on a product that has not yet been released, you cannot be sure it will work in production. What is the impact if the new release does not work? What is the impact if the release is delayed six months? A year? Two years? What is the impact if the product you choose simply will not work? Can you really afford to start again? Can you afford to re-define the problem to match what you've built?

How is the vendor support? You cannot build and support this environment on your own. If you must train the vendor every step of the way, what is the risk of failure?Can you afford to be the first site? There can be strategic advantages to being first, but there is always significant pain and/or risk as well.

You can donwload excellent powerpoint slides on marketing management and business strategy here.

Question 2: Is Your System Mission Critical?
Don't answer this one too quickly. If it's not mission critical today, how quickly will it become critical? Hopefully, you are implementing relationship technologies because there is a business reason to do so, and business issues are mission critical. While many info-structures do not start off as mission critical environments, they very quickly become so important that you have difficulty scheduling preventative maintenance. If your answer to this question is "No," then you need to ask yourself another question: "Why are we building this?"

Question 3: What Is Your Processing Environment?
Do you require iterative processing? It would be difficult to overstate the implications of this response. Iterative processing means an unpredictable environment, and an unpredictable environment means more complex support. For example, in a traditional systems environment, database administrators (DBAs) tune all queries. In the iterative environment, this is simply not feasible.

Your database marketing people will need to take over much of the work you used to rely on DBAs to do. The real value is in the ability to ask unpredictable questions and get a quick response. If the users are asking for the moon, it's probably because they need to get to Saturn to be competitive but do not think Saturn can be reached. They've already lowered their expectations based on past experience—be careful not to lower them any further.

Question 4: How Much Data Do You Have?
This is a tricky question. Don't worry about how much data is on your archive tapes; it's compressed, it's blocked, and it's not all useful data. When you have completed a logical model, how much data is out there to support it? Now, multiply by ten. Why ten? This is a fairly average increase in data volume over the first two to five years of data warehouse or relationship technologies.

When you check the references, be very careful on this point. Vendors like to tell you how much disk space their systems support. More often than not, this indicates the shortcomings of the database managers, not the amazing data volumes they can access. It is not unusual to find systems in excess of 1 terabyte of disk to support less than 100 Gigabytes of real data. Before you know if the solutions can support your data volumes, you must know what your data volumes are.

Consider this question in conjunction with your level of risk tolerance. The largest reference out there for a product got there through much pain. Can you afford to be larger? Can you afford to be as big? Can you even afford to be within 80% of it?

Question 5: How Many Concurrent Users Will There Be?
While this is not as easy to predict as it is in a traditional systems environment, reasonable estimates must be attempted. A system that runs well with one user might turn out to be the world's most expensive personal computer.

In analytical systems it is very common for there to be dozens or even hundreds of users all asking questions (queries) at the same time (when many of these interactions are not pre-defined or pre-coded). The most effective analytical systems provide for any question, at any time, to any data (correlations, as an example), to any user for any potential business opportunities.

In customer relationship systems, such as on the Internet and e-commerce, there is a requirement for more users to access more data in order to be able to make more decisions in an instantaneous timeframe. Be prepared to discuss vast magnitudes of scalability and performance requirements if your system is being opened to your customers and outside users (prospects or channels).

You can download excellent powerpoint slides on Marketing Strategy and Marketing Management HERE.

Source of Reference:
Ronald S. Swift, Accelerating Customer Relationships: Using CRM and Relationship Technologies, Prentice Hall. You can obtain this excellent book here