top of page
SriRam Anjanadri

Master Data Management an essential for Enterprises



We are living in an era where everyone is trying to keep pace with the rate of change. The rate and the volume is so high that data quality issues in any CRM system creeps up at a rate of 2% per month according to the survey of DnB using a global data set.


Enterprises are finding themselves to adapt a pace of change. The rate is so fast that customer data get out of sync in large enterprises having numerous applications to conduct business. This calls for a centralized Data Management solution.


Businesses are usually done with contact person. A Customer, Supplier, and an organization but there is face or human element always involved. The conclusion is that we have two data items. The relationship details between the data items are critical for any business.


Only 30% of relationship can be established with the data captured by internal systems. In this competitive world enterprises are looking beyond the Enterprise data like social media to enrich its data with important attributes like socio-economic & Geo-demographics, Credit Risk, Investment Activity, Financial Holdings for new opportunities and maintaining the current set of customers happy. Infact due to the pace of( change relationship changes) has more probability to hit the social media first rather than the internal systems. For example a job change can easily be traced from linked-in profile. There are two important things firstly termination of a relationship and initiation of new relation which can bring new business.


These social media is public domain data which is available for competitors also. The success factor in this competitive world is to churn the public domain data correctly and enrich the associated internal customer data. The enrichment can fill in important attributes like sphere of influence, formulate new /terminate relationship between existing data items.


Master Data Management (MDM) is an one stop solution for the above a have an competitive edge. MDM is commonly known as “single version of the truth”. We need to observe that it is single instance of information of an entity mainly customer and not data. Any data database tool can remove duplicates but not information. MDM internally has three key workflow Load Master, Matching, Merging and Survivorship. In addition to these as Load master can consume data in any format. The true matching is possible when desperate data sources are standardized. Therefore staging and cleansing are two key preprocessing work flows. In Load masters step loads the standardized data from all downstream sources loaded in a data hub. A master key usually a serial number is generated. In matching two or more record are paired using the master key. This is done by using matching rules created by data analyst or probabilistic matching. In probabilistic matching algorithm pair master keys and assign two integers namely similarity and confidence factor for each or the pairs created. A Single merged record is created from the paired records. A weight factor for each attribute for all the source system is assigned by the data analyst called the survivorship rule. Survivorship rule is used when the single merged record is created from the pairs. A console is used to view the survived record along with relationship and hierarchy. The console is also used to rollback merges.


The ETL tools now have the ability to connect to social media. The MDM solution consumes the social media data as one of the input system to hub while creating the single version of truth. Informatica MDM uses its own tool Power Center to connect to social media.


The MDM is not specific to Customer but for any entity. A MDM solution is essential for any large enterprise to keep pace with the changes in the industry, to have central repository of master which will feed the dimensions of a data warehouse. Additionally the dimensions are enriched from social media to empower data mining/analytics.

0 views0 comments

Comments


bottom of page