Wednesday, November 18, 2009

Evluating Vendor Solutions

First and foremost, determine the functions in your data warehouse that absolutely

need vendor tools and solutions.

 For each type of product you need, carefully list the features that are expected. Divide

the features into groups by importance—high, medium, and low. Use these

groups of features to grade the products you are considering.

 Allocate enough time to research available solutions and vendors thoroughly.

 If you try to incorporate solutions from too many different vendors, you must be

prepared to face serious challenges of incompatibilities and restrictions for integration.

Stay with two or three vendors whose products are most appropriate for your


 Metadata is a key component of a data warehouse. Ensure that the vendor products

you choose can handle metadata satisfactorily.

 The standing and stability of the vendor is equally important as the effectiveness of

the products themselves. Even when the products are suitable for your environment,

if you are concerned about the staying power of the vendor, have second thoughts on

selecting these products.

 Never rely solely on vendor demonstrations as the basis for your selection, nor

should you check only the references furnished by the vendors themselves.

 Test the tools and products in your environment, with subsets of your own data.

 Arrange for both user representatives and IT members of the project team to test the

products, jointly and independently.

 Establish a definitive method for comparing and scoring competing products. You

may devise a point system to score the various features you are looking for in a

product type.

 The success of your data warehouse rides on the end-user tools. Pay special attention

to the choice of the end-user tools. You may compromise a bit on the other

types, but not on the end-user tools.

 Most of your end-users will be new to data warehousing. Good, intuitive and easyto-

use tools go a long way in winning them over.

 Users like tools that seamlessly incorporate online queries, batch reporting, and data

extraction for analysis.

No comments: