Database Answers
  You are not logged in | login
Home Ask a Question Best Practice Careers Contact Us Data Models Search Site Map

Best Practice Manuals  
Below is the Best Practice Manual you selected.

Best Practice for Master Data Management and Reference Data
A Step-by-Step Approach

1.0 Identify Business Champions

These will be vital to your success.
1.1  With Vision

1.2  Running a High-Profile Service or Division

1.3  Having Established a Successful Track-Record

You need Champions who people respect

2.0 Decide the Approach

  • Top-Down, Bottom-Up
  • Ref Data Product Catalogue, CMI, etc..
  • 2.1  Top-Down and/or Bottom-Up

    2.2  POC or Feasibility Study

  • Choose either a Proof-of-Concept or a paper Feasibility Study.
  • You could hold Design Reviews
  • 2.3  Management Involvement

  • You must get buy-in at the 'working level' from key managers.
  • 2.4  Decide your Success Criteria

  • You must establish how your success will be measured, preferably on a regular basis and not right at the end.
  • 2.5  Define the Enterprise Data Platform

  • For example, start with this one from DatabaseAnswers.
  • 3.0 Determine the Standards

  • These can be Internal, External, National or International
  • 3.1  Easy where defined

  • 3.2  Look for obvious Data Leaders

  • eg Social Services for Ethnic Origins
  • 3.3  Create a Glossary for Data Mapping

  • You need a Glossary from early on.
  • It provides a structure, signs of progress and helps people feel involved.
  • 3.4  Aim for Buy-In

  • It's difficult but very important to get buy-in early on.
  • Remember this is a business Project, not a technical one.
  • 4.0 Perform a Data Quality Audit

  • Design the Infrastructure
  • Choose Software Tools
  • Establish the Governance
  • 4.1  Sell the Importance

  • Especially to management
  • 4.2  Prepare a Business Case

  • Boring but essential.
  • You must look at your Project thru the eyes of the Sponsors.
  • 4.3  Carry out Enterprise-wide DQ Audit

  • Tools and Techniques are Enterprise-wide
  • 4.4  Data Profiles suggest Standards

  • A simple example is an analysis of Genders.
  • We might expect to get many examples of M/F and a few 0/1's so we would choose M/F.
  • 4.5  Obtain Buy-In from Data Stewards

  • It's good to find people who will take ownership of different kinds of data.
  • They then feel committed to making things work.
  • This reduces your workload and increases your chances of success.
  • If you don't get buy-in, your job is going to be very difficult.
  • 5.0 Define the Scope

    This is where we establish the 'Things of Interest' that are in Scope.
    5.1  Reference Data

    For example, Payment Methods.
    5.2  Products

    Choose the most appropriate Data Model, such as Product Catalogs.
    5.3  Suppliers

    If Suppliers are in scope, choose the most appropriate Data Model, such as B2C e-Commerce or Customers and Campaigns.
    5.4  Organisation Structure

    If the Organisation Structure is in scope, choose the most appropriate Data Model, such as Hierarchies or this one for Organisations operating in different locations.
    5.5  Customers

    Probably the single most important 'Thing of Interest'. This Web Site offers about 25 to choose from.
    5.5.1  Commercial and Personal Customers

    This Model uses Inheritance to represent the differences between Commercial and Personal Customers.
    5.5.2  Customers and Products

    This Model is a simple starting-point and the first step would be to remove Products.
    Home Ask a Question Best Practice Careers Contact Us Data Models Search Site Map