THE HOLY GRAIL: DETERMINING DATA WAREHOUSING RETURN ON INVESTMENT
by Jim Gallagher
According to the Meta Group, the average data warehouse takes 3 years to build and costs 2-3 million dollars. With this kind of investment in time and money it is naturally important to review what the return on this investment is expected to be. And even if you are building a "data mart", or a more modest warehousing effort, it is important to tie it back to business value. Anyone who has gone to a CFO for funding knows that.
It is quite obvious that a direct involvement (and sponsorship) from your company's business community is key to the success of a warehousing effort. But that is the same for almost any IT system with business users. The difference is that in building an operational system, it is relatively simple to measure the reduction in cost to produce an invoice, or the ability to process billing faster, or the savings gained from reducing the downtime of an old, overworked system by replacing it with a new one. When looking at decision support applications running against a data warehouse, it becomes more difficult to quantify the financial impact in advance.
Experts will tell you that most successfully implemented data warehouses have a tremendous return on investment. One IDC study mentions returns of over 1000%. But ultimately, that return is not really known in concrete terms until the warehouse is actually being used. That does not help the executive who is trying to get funding approved, and is dealing with someone who wants to make sure the return on investment is real before granting the funds.
So how do you do it? Unfortunately there is no "silver bullet" algorithm or answer that can be generically applied. But there is a process that you can go through that will allow you to predict gains.
The key is to maintain a focus in business terms, and in laying out guidelines that can be used to measure an improvement in the business processes/events the warehouse is intended to address. This means looking at Key Performance Indicators your warehouse is ultimately addressing, and working backward through the processes that manage them. An example would be to look at Customer Retention, and identify how that performance measure is tracked. One technique that works well is to construct a "day in the life" scenario by interviewing key business users involved in the process. If properly facilitated, these interviews can identify opportunities that the data warehouse can address.
In the case of customer retention, it could be determined that certain information needed to solve a customer problem takes a week to derive using existing systems and methods. This of course may contribute to the loss of a customer. If the warehouse can provide that information faster (in a few minutes) a direct impact can be assessed. This is a very simplistic example, but the process to do this type of analysis can be implemented on a larger scale to create some fairly specific examples of the business impact of the warehouse. Quite often, the best "nuggets" of ROI for the warehouse are hidden in the minds of business people who have no idea that they are holding the key to tremendous opportunities. Why? Perhaps because they never had the advantage of decision support in the past, and are not aware of exactly how powerful it can be. The interviewing process can help uncover these "nuggets".
The additional benefit is the fact that you will be interacting with a business user and identifying their areas of "pain". This will create opportunities for you to address the areas of pain, and deliver real value and credibility for the warehouse.
Another way to gather some ROI information is to talk to companies (especially those in the same business as yours), who have already implemented warehouses. If you can get information on how their warehouse has "paid for itself" you may be able to draw direct correlations to your own initiative. Research organizations such as IDC, the Meta Group, and the Gartner Group may also have some examples that can be useful.
There are some general business benefits that could be expected, which create a starting point for your ROI:
There are also some IT based benefits that a data warehouse can address, such as turnaround on reports, a reduction in workload to create custom reports, and a general increase in the quality and standardization of data across the subject areas affected.
The other key to making sure your data warehouse pays for itself is to pay attention to the "non-technology" aspects of the project. This includes addressing user training, cultural issues, and related business process issues. And it is important to make sure your knowledgeable "power users" have the power to act (make decisions and implement them) on the information that they derive from the warehousing applications.