DS0509_Opinions_p12-p13_Trillium:Layout 1 29/4/09 17:25 Page 12
DATA QUALITY
A guide to
implementing
data quality
Planning a step-by-step approach, based
tools or those under consideration can
support data quality process execution on all
on a method for engaging process, people
platforms of your source and target systems.
Better still, consider your organisation’s long-
and technology will help ensure success, says
term vision and ensure that your product
selection will extend later to other
Ed Wrazen of Trillium Software.
environments in your enterprise.
PHASE TWO: BLUEPRINT
Now prepare plans and detailed designs to
meet project requirements, while mitigating the
ncreasing numbers of organisations is derived from both proper ‘Syntax’ (data risks discovered in phase one. It’s essential too
I
are running data quality initiatives to format and structure) and ‘Context’ (the to devise a communications plan, designed to
optimise the value of their customer, meaning of data to the business). Data stewards inform as well as engage the business.
product, supplier, financial and credit should hold a good understanding of the data Next define the data quality metrics you’ll
risk data – right across the enterprise. assets in their business domain, business rules track: such as how many records are non-
To meet both your business and that should be supported by the data, how that compliant with key business processes, support
technical goals, it’s important to follow a well- data supports current and planned business for defined data standards, rules and business
defined methodology. The guidance notes processes, issues that have arisen and manual objectives, how many incomplete records you
below are an extract from the Trillium Software corrections that have been made. have and suspected duplicates. Data profiling
white paper “Data Quality Essentials: for any The group will need to scope the project. Clear and discovery tools and exception reporting
Data-Intensive Project” available at parameters are needed around the data you are allows data management, project and business
www.trilliumsoftware.com capturing, moving, cleansing, standardising, team members to test for such suspected issues
matching and deduplicating and enriching, and and define the baseline. Better tools will even
PHASE ONE: PROJECT PREPARATION its use. You must determine whether the identify issues you didn’t know to look for.
In this phase you will evaluate business suggested data actually exists and where, its It will be necessary to take a deep dive into
objectives and assess resource requirements. current quality and whether issues within it can data extracts, representative of the actual data
Define your team members and their actually be cost-effectively resolved at all. A data that will be used as part of the production
responsibilities, the scope, expectations, and profiling and discovery tool will assist all the system. The purpose here is to understand
deliverables of the project and conduct an above activity and enable a data quality risk what mappings, transformations, processing,
analysis of the current state of your data. Define audit and assessment to be undertaken. cleansing and rules must be established. It
issues and risks. Produce an outline business Identify any current data quality technology really is important to collaborate with the
case for presentation to management. in place. It may be that other divisions or business in the analysis of source data and to
Your team should include ‘data stewards’, as departments already have a solution, however, reach a consensus on what should be done.
subject matter experts from affected business existing tools will often be point solutions, However, many technical project managers
areas, as well as IT. This is essential because without regard to the entire enterprise. As find this time consuming and frustrating as
successful interpretation and treatment of data such you’ll need to ensure that either existing business users often lack the technical skills to
12 OPINIONS ON DATA | MAY 2009
Page 1  |  Page 2  |  Page 3  |  Page 4  |  Page 5  |  Page 6  |  Page 7  |  Page 8  |  Page 9  |  Page 10  |  Page 11  |  Page 12  |  Page 13  |  Page 14  |  Page 15  |  Page 16  |  Page 17