How to Achieve a Single Customer View



Similar documents
Understanding Data De-duplication. Data Quality Automation - Providing a cornerstone to your Master Data Management (MDM) strategy

Chimpegration for The Raiser s Edge

Blue Cannon Lead Generation

Creating an Excel Database for a Mail Merge on a PC. Excel Spreadsheet Mail Merge. 0 of 8 Mail merge (PC)

Click-n-Print User Guide

Best Practices for Creating and Maintaining a Clean Database. An Experian QAS White Paper

AMB-PDM Overview v6.0.5

NHS Central Register (NHSCR) Governance Board Meeting 31 st July 2012

Everyone deals with the problem of maintaining master data. To illustrate, consider the list of contacts on your smart phone

Blue Cannon Lead Generation for IFA s

6 Steps to Creating a Successful Marketing Database

How To Buy B2B Data. A clean, accurately targeted and relevant list is a critical factor in the success of any marketing campaign.

Data.com Record Matching in Salesforce

List Cleansing: Questions Answered

Experian Data. A simple insight into our solutions. Experian Data Quality Tools

Welcome to the webinar Does your department or company use the valuable data it collects to plan for future needs and trends?

National Child Measurement Programme 2015/16. IT System User Guide Part 3. Pupil Data Management

Data counts. Addresses. Job Number: / Dear Mrs Test

Table of Contents. Copyright Symphonic Source, Inc. All rights reserved. Salesforce is a registered trademark of salesforce.

July Background

Chapter 16: Follow-up

Online RESTful API documentation. V-Series 5.1

Technical implementation of multi channel content management

Background Who are AddressWorks? How do I get a Statement of Accuracy? Data Cleansing Frequently asked Questions...

Data Products and Services. The one-stop-shop for all your business-to-consumer data requirements

The Cost of Duplicate Data in Enterprise Content Management

The Essentials of Finding the Distinct, Unique, and Duplicate Values in Your Data

List Hygiene Products & Services Detailed Overview

DonorCommunity CRM UPLOAD GUIDE. Version 1.0. DonorCommunity, Inc NW 8 th Street Suite No. 205 Sunrise, FL DonorCommunity, Inc.

Data Solutions - Reducing Performance Loss

Manager/Supervisor & OHS Consultant User Guide

Contactegration for The Raiser s Edge

Data Imported and Displayed

Section of DBMS Selection & Evaluation Questionnaire

TUTORIAL: Campaigns Gold-Vision 6

Employment intermediaries: data requirements for software developers

ESKIDMS3 Database management software

Foundations of Business Intelligence: Databases and Information Management

OpenInsight Data Encryption at Rest (RTIDER)

IAB/ABC International Spiders & Bots List

Data Cleansing and Maximizer

Understanding DSO (DataStore Object) Part 1: Standard DSO

ACHIEVING YOUR SINGLE CUSTOMER VIEW

Business Telephone Banking Administration form

Payroll Time Clock Import - Quick Start Instructions

Creating Tables ACCESS. Normalisation Techniques

Manage Workflows. Workflows and Workflow Actions

How to Manage and Organise Shared Drives. Guidance for Administrators

Internal Control Deliverables. For. System Development Projects

Direct Marketing (Customer Relationship Management) Blackbaud Europe, Ltd

Chapter 6 FOUNDATIONS OF BUSINESS INTELLIGENCE: DATABASES AND INFORMATION MANAGEMENT Learning Objectives

Lead Management User Guide

Insurance Training 101

Data Integration Alternatives Managing Value and Quality

Session 16 Standard Student Identification Method

itg CloudBase is a suite of fully managed Hybrid & Private Cloud Services ready to support your business onwards and upwards into the future.

Course MIS. Foundations of Business Intelligence

Deutsche Post Address Group

Creating an Excel Spreadsheet for Mail Merge. Excel Spreadsheet Mail Merge. 1 of 9 Design & Print Offline: Mail Merge

THE SIX PRIMARY DIMENSIONS FOR DATA QUALITY ASSESSMENT

ENTERPRISE CONTACT MANAGEMENT

The Real Challenges of Configuration Management

Can good data deliver a be er customer experience? Discussion Paper

Ad Hoc Advanced Table of Contents

How To Improve Your Mailman

Demographic Batch Search (DBS)

HIGH PRECISION MATCHING AT THE HEART OF MASTER DATA MANAGEMENT

Despatch Manager Online

Marketing. o/ . TechStorm NORTH DALLAS PARKWAY SUITE 125 ADDISON, TX,

How to: Account for Settlement Discount VAT Rule Changes from 1 st of April 2015

MAXMAILER USER GUIDE

Estimating and Vendor Quotes An Estimating and Vendor Quote Workflow Guide

Best Practice Data Collection for Marketers

CUSTOMER MASTER DATA MANAGEMENT PROCESS INTEGRATION PACK

Business Operations. Module Db. Capita s Combined Offer for Business & Enforcement Operations delivers many overarching benefits for TfL:

Data Quality Improvement and the Open Mapping Tools

MXSAVE XMLRPC Web Service Guide. Last Revision: 6/14/2012

Policy Document Control Page

Demystifying Big Data. James Rawlins Senior Consultant

Table of Contents. KB Group Suite KB Bulk Mail version 1.0 Page 2 of 36 Copyright KB Group (UK) Ltd

Your Guide to Kelly Web Time

MICROSOFT DYNAMICS CRM Roadmap. Release Preview Guide. Q Service Update. Updated: August, 2011

Master Data Management

enicq 5 External Data Interface User s Guide

Access Coordinator (AC) User Guide. September Customer Registration Service (CRS)

Gift Aid Charities Online - Schedule Spreadsheets Information Sheet

Cathay Business Online Banking

UK Data Archive Data Dictionary

Gardners ebooks frequently asked questions

RETAIL Sage Manual (Retail Point of Sale and Customer Management System) Version 1.2

CUSTOMER MASTER DATA MANAGEMENT PROCESS INTEGRATION PACK

A Simplified Framework for Data Cleaning and Information Retrieval in Multiple Data Source Problems

ing a large amount of recipients

PROCESSING & MANAGEMENT OF INBOUND TRANSACTIONAL CONTENT

Mail & List Segmentation. For the BC Blackbaud Users Group Tamara Wojdylo & Chelsey Ireland VGH & UBC Hospital Foundation

Education Module for Health Record Practice. Module 2 - Patient Identification, Registration and the Master Patient Index

Marketo Integration Setup Guide

Batch Processing Version 2.0 Revision Date: April 29, 2014

What is Data Virtualization?

Transcription:

How to Achieve a Single Customer View 1.0 Introduction Clients want to obtain a Single Customer View of their contact database/crm system to let them understand the types of individuals/businesses that they are dealing with. Contact data is held on a variety of databases, the quality of the data entered will vary and as we have seen from experience, customers may appear in more than one database leading to multiple communications and wasted money; or alternatively they may have moved address, again wasting communications. Not only is there a cost for holding all this incorrect data it also leads to a distorted picture of the members on the database e.g. if demographic profiling is used the results will show too many customers and possibly skew the profiles. 2.0 The Business Case Companies need to know how many customers/supporters/members they have, they want to cleanse the data to save money on any marketing campaigns. Importantly they also need to be assured that they are not in breach of the Data Protection Act for, Personal data shall be accurate and, where necessary, kept up to date. The cost of handling returns and customer complaints can also be reduced. 3.0 The Problem Statement Companies have many databases containing details of customers/supporters/members. The chances are that people could appear in more than one database and in some cases all of them. The task is to merge the databases to present a unified view. Once the initial merge is completed then it is important to implement a process to maintain the database by regularly cleansing the data. Step 1 Agreeing Rules A set of rules and processes need to be agreed on how to consolidate certain records and identify others for further examination. This is to ensure that important records are kept and only to archive duplicate or inaccurate data. Once the specification are agreed and signed off, the rules and techniques detailed below can be applied to create a programme to speed up the identification and merging of duplicate records. Step2 Data cleansing and validation rules. It is only possible to move onto de-duplication once the data cleansing is completed, prior to this the addresses would be in a non standard format and it would be difficult to find any matches at all. Standardising the data maximises the chances of finding a match. How to Obtain a Single Customer View Page 1

The cleansing procedures: 2.1 PAF Cleansing This matches addresses on the databases against the Postcode Address File (PAF) from the Royal Mail, automatically correcting errors where possible. 2.2 Telephone Validation This initial validation checks if a number is callable based on factors such as, number of digits and dialing codes etc. If required there are more detailed validation methods to confirm that there is a live line. 2.3 Email Validation Initially this will be validated at server level. This means that data8 checked that the mail server for the domain is alive. Again there are more detailed checks that can be applied. The levels of validation are chosen dependent on what the client wishes to achieve and the rules agreed. Step 3 De-duplication Once the data is cleansed it is possible to look for duplicates, in order to identify duplicates certain rules and assumptions needed to be made, based on this, bespoke algorithms are designed to produce the required results. Candidate Retrieval This algorithm works by considering each record in the database and finding any suitable candidates to be considered as duplicates with the currently considered record. The candidate retrieval process can be: a) Records with the same Name and Address (where populated) b) Records with the same Name and Date of Birth (where D of B populated) c) Records with same Name and Email (where email populated) The process then becomes recursive and each new candidate gets considered to build a full possible chain of records that are similar and may be considered as duplicates. How to Obtain a Single Customer View Page 2

Fuzzy Matching & Formatting The quality of data capture across the databases will vary, therefore it is necessary to develop a custom set of fuzzy logic matching. See table below. Field DOB: Address: Email: Allowable Difference Differs by one corresponding digit 1st of the month date to date with same month and year US date format to UK. i.e. MM/DD/YYYY to DD/MM/YYYY Data8 PAF matching Differs by one character Just using exact matching would allow duplicates on the database, small differences in the matching routine can be allowed. For example is John Smith the same as J Smith, if they are, the records should be merged. Exact matching would identify them as two different records, doubling the marketing communications that John Smith would receive. Telephone: Data8 Telephone Formatting Identification of Duplicates data8 then create a candidate list of potentially similar records, and develops a set of rules which are used to break the list into three categories SAME - Records identified as the same and therefore merged MANUAL - Records identified as possibly the same but a significant difference indicates a manual check must be performed. UNCHANGED - The remainder rejected from the duplication process and returned unchanged. The rules with which data8 allocated candidates vary here is an example of a recent de-duplication exercise: 1) SAME The following details must match all records in this chain: Name must be the same (i.e. the name has to be populated and match at initial level) Non conflicting Address (Allow blanks) Non conflicting Date of Birth (Allow blanks) Over18 flag must be the same (Allow blanks) 2) MANUAL In the following situations, all records that meet the below criteria were put into the manual investigation pot: Name must be the same Addresses will conflict (Not blank) How to Obtain a Single Customer View Page 3

Non conflicting Date of Birth (Allow blanks) OR Name must be the same Non conflicting address (Allow blanks) Conflicting Date of Birth (Not blank) 3) UNCHANGED Any records that are not matched in any of the above criteria were put into the Unchanged pot and returned without merging. Step 4 Merge Rules Once candidates are identified as members of the SAME duplicate set, then rules to merge fields have to be defined. What is illustrated is part of the merge priority of the fields. Taking the merging of the Firstname field as an example. Firstname John Jon Jonathon Jon This example assumes that the surname and address appears seven times in the databases, however the first name associated with surname and address varies and in some cases it is blank. The choice therefore is to select from John, Jon or Jonathon, blank values are ignored. The value to take would be Jon using the Winner Takes All rule, it appears twice while John and Jonathon appear only once. Step 5 Output Files Two output files are generated to show merged record details and a further file where there were possible duplicates requiring manual investigation. After a manual intervention any duplicates are removed and the files merged. 1. The first output file contained the resultant merged and unchanged records based on the previous rules. A column will also show any related records which contributed to the merge process. 2. Another output file is created which shows all records merged and any associated records. How to Obtain a Single Customer View Page 4

3. A third output file shows records which were identified for manual investigation. This includes a field showing all related records. After checking the files requiring a manual check the data can be then all brought together as one database. Step 6 Goneaways, Movers and Deceased Once all the data is one database it is then time to identify those people who have gone away and where possible find an alternative address i.e. movers. Also it is important to identify those who have died to avoid the embarrassment of sending out a marketing communication. Step 7 Maintaining the data In the longer term the data should be cleansed very couple of months, the costs will be minimal as most of it is already correct. Alternatively by using web api it is possible to validate the contact data, in real time, as it flows into the company, confirming address, telephone number and email as it is input on a website or any business application including POS. How to Obtain a Single Customer View Page 5