-
-
Marketplace
-
Channel Resources
Articles from this Site
-
One brand, one Web site! DM Review is now the home of all the content you're used to at BIReview.com and much more. If you are registered at BIReview.com, you're already registered at DM Review. If not, take just a moment to sign up for all the free services we have for you at the new DMReview.com.
Lessons Learned
Business intelligence (BI) is a broad collection of applications and technologies used to analyze data and empower individuals to make better-informed business decisions. Companies have moved the purchase of BI tools to the top of their priority lists as they recognize the benefits these tools can offer in their efforts to lower costs and improve productivity. To effectively choose the optimal BI product for the firm it is critical that both IT and their business counterparts work together to understand the technology as well as the business issues and goals. Focusing on technology or business issues independently of each other will often lead to failed product evaluations and/or implementations.
Choosing the right BI product suite can be a daunting task as each vendor's product offers similar features and functionality but in different ways and at varying price points. Having been involved in a number of these evaluations at Merrill Lynch, I have found that the ultimate decision on which vendor to choose is not an easy one and the factors for choosing the BI tool suite will vary by firm. What I would like to leave you with in this article are some of the lessons that I have learned from my trials and tribulations associated with evaluating and ultimately choosing BI tool suites.
Lesson 1
It is all about putting data to practical use! As technologists, we have a lot of fun discussing and evaluating the tool suites. Business users however, just want to see their data in the most straightforward form. Three-dimensional rotating graphs are fun to see and play around with but most users will never use them.
Lesson 2
Know the requirements of the users before you start evaluating the tool suite. Collect sample reports or report mockups. The client must know exactly what they want to accomplish with the tool or this will be an open-ended evaluation. Develop a scorecard with the clients so that they can easily compare the products. For example, items such as single sign-on, role based security, support for Excel, visualization offerings (pie, bar, line, gauges and so on), and event scheduling can be listed on a one-page scorecard. These items can be grouped into categories and graded on a scale from one to ten. Having a predefined scorecard with numerical ratings provides a communication vehicle that can be used by both the technology and the business areas to rate the products. In addition, make sure that all technical terms such as publish and subscribe, ODBC access, multi dimensional analysis (OLAP, ROLAP) are clearly defined and understood in language relevant to all parties so that it is an inclusive and not an exclusive evaluation.
Some end users are not familiar with the technology nor do they want to be. Therefore you must segment users into categories such as novice, intermediate and power user. Make sure you devote enough time with each user group and demonstrate how the BI suite can help them perform their jobs better. Take the time to solicit and catalog end user input and suggestions. Just having management backing does not mean you automatically will gain acceptance from the business community; you must earn their trust and approval. If you do not gain their acceptance they will continue to use their old tools and methods and will be resistant to the BI rollout.
Lesson 3
Do not try and be nice to all vendors and evaluate every tool in existence. Going to lunch or dinner with every vendor that wants to explain the virtues of their product suite will extend the evaluation considerably and perhaps your waistline as well. The maximum number of tools that you should evaluate should be three. Your filtering criteria could be vendor market size, global support, functionality, an existing relationship, or other limiting criteria.
The risk of vendor viability and ongoing product support is a key part of the decision-making process. The BI marketplace has experienced significant growth over the last few years as well as significant consolidation as each vendor has tried to complete their BI product suite by merger and acquisition. For example, Business Objects purchased Crystal Reports, Hyperion purchased Brio, and IBM purchased Alphablox and Ascential. Make sure the BI vendor has a viable long term strategic vision. You never want to be the only customer using a product.
Lesson 4
Creating a proof of concept (POC) is much better than performing paper evaluations. Think about sending the vendor your data (making sure that the vendor is willing to sign your firm's non disclosure agreement) and let them use their resources (hardware, software and human) at their site to create the POC. The POC can be accessed via the Web and you will never have to leave the confines of your firm's building. Remote access via the Web has been improved over the years and can expedite the evaluation process dramatically. By performing the POC in this manner you can leverage the vendor's resources to do all the administration and set up and your team can focus on gaining insights into solving business problems. Also, a rapid POC allows the client to see their data quicker (see lesson number 1) and they can see firsthand the benefits the BI tool suite will offer them. Finally, let the clients view each vendor's tool suite in a side-by-side comparison lab.
Lesson 5
Don't get stuck with "SHELF WARE!" Buying a complete BI suite or purchasing complementary products from the vendor at the same time can lead to shelf ware. Think about how much of the Microsoft Office suite you have paid for, and how much of it you really use. You need to stay focused and buy only the products that satisfy the requirements, nothing more and nothing less. The vendor will be more than willing to sell you additional products at a later time. Also, try to negotiate pricing with the vendor so you can buy licensees in small increments - such as 25 to 50 - rather than large blocks. Buying a larger block of licenses may sound like the right thing to do but remember that initial estimates of use are usually overstated.
Lesson 6
Single sign-on is always a top priority to clients. They want to sign in once and have access to the entire BI product suite. Personalization based on their unique ID is also a key requirement. This can be satisfied sometimes within the BI product itself or may require an additional portal layer. Remember, most users are quite familiar with the popular Internet portals such as my.Yahoo.com and msn.com and want similar functionality and features.
Lesson 7
Metadata about the data that the BI suite is using is crucial to the success of the BI rollout. Clear, concise and easily accessible definitions about the data attributes will promote the use of the BI tool suite as it will empower the end user to create their own reports. In addition, formalized training should be given on a regular basis to continue to promote and advertise the best ways to use the BI suite. Failure to provide the proper amount of metadata and tool training will result in frustrated users and an expensive, underutilized product suite.
Lesson 8
One tool cannot satisfy all clients. Do not be afraid to augment the BI tool suite with other tools required by the user or lacking from the BI suite. For example, not many BI suites come with heatmaps, and in my personal experience I have found heatmaps are often an excellent way of presenting key information to the client. Also, do not forget the power of Excel. Financial officers are extremely adept at using Excel and usually balk at using anything else.
Lesson 9
Just as a lawyer will never ask a question he or she doesn't know the answer to, you must know the underlying data that you will be using in your BI evaluations. Don't use the sample data that comes with the BI product, use data from your production applications. This will allow the clients to relate to their data and envision how the BI product can help them solve their everyday problems. In addition, make sure that the data you choose does not have any quality problems. A data source that displays 999 states in the United States region table will be exposed and exacerbated by the BI product.
Lesson 10
Try to limit the evaluation period to a fixed time frame. From my experience, three to six months is enough time to install, test and choose the BI product suite. Don't follow the trends, pick the correct product for your organization and implement it. Long evaluation periods generally lead to scope creep, BI product changes (note the recent M&A activity), management sponsor changes and/or budget issues that could derail the project.
Conclusion
It is important that technologists understand the business goals of their firms and use their technology know how to uncover the product that will allow the business areas to gain insights and achieve their business objectives. Conversely, business users need to partner with IT and accept that some technology tool and data training can greatly improve the BI selection process and their overall success using the BI tool suite. Though some lessons can only be learned the hard way, following these steps can help organizations improve the process of identifying, evaluating and purchasing a BI suite.
Howard Goldberg is a vice president at Merrill Lynch with more than 22 years experience, primarily in databases and data warehousing. He can be contacted at howard-goldberg@ml.com .
Howard Goldberg is a vice president at Merrill Lynch with more than 22 years experience, primarily in databases and data warehousing. He can be contacted at howard-goldberg@ml.com .
For more information on related topics, visit the following channels:
