Following from the previous blog post, this post will explain the next four steps in the process for implementing MDM within an organisation.

Implement a data governance program and data governance council

The data governance council should be assembled from people who know the data and have the authority to decide to decide what should be stored in the master data, for how long, authorise changes to be made to any of the master data etc.

Develop the master data model

In this step, you decide what attributes should form part of the master data, what data types, restricted values lists, default values, data types etc. Also create a mapping sheet from the source systems to the MDM system which will explain how the attributes can be later on translated during the implementation phase. It is worth noting note to aim to please every stakeholder in this process otherwise you will end up with a large set of master data attributes making the master data very complex.

Choose a MDM tool

Select a MDM tool which will help to collect, transform, standardise, match, link and merge data and then store it into the MDM data hub. You may also need to purchase additional tools for maintaining and using the master data.

You can either select a single tool which is capable of handling multiple domains or choose separate tools i.e. best-of-breed. Even though there are multi-domain MDM tools available on the market, it is probably best to go for the best-of-breed MDM tools especially if the requirements are complex.

When selecting a tool, ensure that it has the capability to identify and resolve data quality issues. In addition, it should have hierarchy management and versioning/history features.

Design the infrastructure

Once you have configured the MDM tool, it is time to set up processes which will use and maintain the master data after the MDM solution has gone live. This may involve changing existing systems and IT infrastructure. The infrastructure will have a dependency on the MDM system and therefore reliability and scalability are additional factors to take into account when deciding on which MDM tool to purchase.

So these were a few more steps involved in the process. The remaining steps will be explained in the next blog post which will be published in the next few days.

This blog post and the next two posts, which will be published in a few days, describe the process of implementing MDM within an organisation.

The order of steps used to describe the process in these blog posts should not be treated like a recipe where you  have to follow it in a certain order but the steps can be implemented in a slightly varying order. Obviously, there are steps which must be done before the other but where there is flexibility available in ordering some steps, should be apparent from the detailed descriptions.

This is how to implement MDM:

Identify the sources of master data

Begin with identifying where the master data is stored currently. This can be quite a revealing exercise, as it often amazes organisations to find out in how many source systems master data is stored.

Identify the producers and consumers of master data

Next find out which systems create and use the master data. Finding out about which systems create master data is usually easy, but finding out about the systems which use the master data can be more difficult.

Collect and analyse the meta data for the master data

Gather information about the entities, attributes and their definitions which will be part of the master data. This also includes finding out about the default values attributes use, restrictions, relationships, who is responsible for the definition and maintenance of data attributes etc. It is important but can be difficult to know who is responsible for the data attributes and maintenance. If all the meta data is stored in a repository then this should be quite a straightforward task, but if you have to go to database and source code level then this can be a very time-consuming and painstaking task.

Gather the data stewards

These are the people who know about the current data and will be in the best position to advise how to transform the current data into master data. Data stewards should include people who know the data, the architect who will be involved with the MDM architecture and representatives from the business who will be prospective users of the master data.

These are just a few steps in the process so far. Some more steps will be explained in the next blog post which will be published in the next few days.

ETL (Extract Transform Load) is mainly used for Business Intelligence and Data Warehouses. It is mainly used for moving data from one platform or system to another also known as data migration. Within ETL tools you can configure the sources of data, how it should be transformed, standardised, matched and merged and finally loaded into a single data hub. The data is only as good as the last ETL batch job which loaded/updated the data into the data hub.

On the other hand, MDM tools contains ETL functionality but offer a lot more than just data migration. In addition, MDM tools offer versioning and auditing, hierarchy management, workflows etc. MDM is also about changing some of the existing business processes if needed and introducing data governance. MDM is not about moving data from A to B, but about real-time or near real-time information retrieval. In addition, the data is also kept up-to-date real-time or near real-time. MDM’s purpose is to provide a ‘single version of the truth’ or also described as to serve as a source of ‘golden records’.

This is the last blog post as part of the blog series explaining the top five influencing factors on a MDM project plan. In this post, it will be described how the size of the MDM problem influences a MDM project plan.

The size of the current MDM problem is related and similar to the requirements and priorities factors. During the initial discussions and analysis across the different departments within an organisation the interviewees and stakeholders may point out various issues which helps to identify all the current pain points. After the analysis you should get a good picture how big the problem is. Based on this, you may want to put some of those problems at a higher priority to deal with them first. The project plan needs to take this into consideration.

Furthermore, the size of the problem may also mean that the organisation wants the MDM consultant to resolve especially the major issues as soon as possible. So this information is something that the MDM consultant should share with the Project Manager so that he/she can devise the MDM project plan accordingly (if possible).

Depending on the size of the problem, this could also mean a longer time to implement MDM within the organisation. Similarly, it could mean that more resources may be needed – either throughout the entire MDM implementation or more likely at certain phases of the project.

With this said, this ends here the series of blog posts explaining the top five influencing factors of a MDM project plan.

The next aspect which influences a MDM project plan are the priorities an organisation has.

A MDM consultant and project manager may create a project plan which will incorporate best practices of implementing MDM solutions. Those best practices may suggest a specific order in which the MDM solution should be implemented. However, if the organisation has priorities then the project plan may need to be created accordingly.

For instance, a business which is not familiar with the concepts of MDM may first want to test the water by trying a small MDM pilot project. In this case, the MDM project plan should include this pilot project.

Similarly, various deliverables will be made as part of the MDM solution. If there are other projects being implemented which need the master data, then the project plan must take this into account and the MDM delivery time lines must be aligned with the projects’ time lines.

These are just some examples how priorities from an organisation may influence the MDM project plan. There are others which could be added to the list, but this gives a few examples of those priorities.

In the next and last post of this series of blog posts the last factor influencing MDM project plans will be explained; size of the problem. The blog post will be published in the next few days. So watch the space!

 

Following from the previous two blog posts, this post will explain the next factor which influences a MDM project plan; the requirements.

It is self-explanatory that requirements influence a MDM project plan – and any other type of project plan. However, what is worth noting is to be aware of anything that may impact the MDM project or requirements included in the plan.

This could include future projects. If so, it is important to find out essential details about those projects such as their purpose, architectural impacts, what impacts it will have on the MDM system, what data it will be using, project time lines etc. All this will help to align the MDM project plan with such other projects in the pipeline.

In the next few days the fourth blog post will be published as part of the series explaining the five factors impacting MDM project plans. The next blog post will explain how an organisation’s priorities influence MDM project plans.

Following on from the previous blog post, this post will explain how constraints on resources impact MDM project plans.

Resource constraints are similar to time constraints and therefore cost organisations money. Managers within the organisation may think that resources should be employed on tasks to maximise the value they get out of paying employees a monthly salary. Hence resources may only be made available to contribute to the MDM project for a short period of time or only part-time basis. This can be challenging and must be taken into consideration in the MDM project plan particularly if those resources are valuable to the execution of the MDM implementation.

Related to the above point and assuming that an organisation does have resources available for the MDM implementation, the next question is whether the resources have the right set of skills that are required for the MDM project. If not, then new resources must be recruited who do have the relevant set of skills needed for the MDM project.

Therefore when creating the MDM project plan it is important to be aware of any such constraints on resources. There  are obviously other questions around resource constraints, but this blog post only explained the most common challenges around resourcing when drawing up the project plan for a MDM implementation.

In the next few days another blog post will be published as part of the series explaining the five factors impacting MDM project plans. The next blog post will explain how requirements influence MDM project plans.

The diagram on the left shows the top five factors which influence a MDM project plan:

  • Time
  • Resources
  • Requirements
  • Priorities
  • Problem Size

Each of these factors will be explained in detail as part of a series of blog posts over the next new few weeks. So watch this space!

To begin with, this blog post will explain in detail how a MDM project plan is influenced by time constraints.

Like any other IT project plan, MDM project plans are also influenced by time constraints. A MDM implementation usually takes anywhere from six months to two years with some implementations exceeding this time frame occasionally.

If you are putting together the project plan you may derive a plan which goes beyond the two year time limit. This may be for very legitimate reasons. However, to be able to convince the business sponsors to invest into the MDM initiative and bring them on board, it is best to create a MDM project plan which does not exceed beyond two years for several obvious reasons. One of the reasons is that it implies more costs associated with the MDM project. Another reason is that depending on the size of the company, a number of other projects may be in the process of being implemented or planned for the near future which will be impacted by the MDM solution. In order to keep pace with the evolution of the IT infratructure – both software and hardware – within the organisation, a two year time period is possible to plan for and is predictable. Predicting what projects and infrastructure changes will take place beyond this is not impossible, but is more difficult.

In case the project does go beyond two years it is probably a good idea to use some project management strategies to keep the implementation time frame below two years. This can be done for instance by increasing the number of resources to work on the project, applying best practices and agreeing with the business on ways of working to gain maximum efficiencies.

So this was an explanation how time influences the MDM project plan. In the next few days another blog post will be published which will discuss how resource constraints influence MDM project plans.

I posted a blog post earlier this year about the definition of MDM, but thought that I add some more notes to it through an additional blog post.

MDM as already explained in one of my previous posts includes the tools and processes to create and maintain master data lists. Even though MDM needs technology to function, it often is more about the business and involving them to make it a real success. Essentially, both are mutually exclusive.

As mentioned, MDM is about the creation and maintenance of master data. Therefore simply creating a master data list is not enough. It is vital that tools and processes are put in place to maintain the master data for future use. MDM costs a lot of time and money and it would all be wasted unless it is properly maintained.

If an organisation has not implemented an MDM solution before, then it is best to run a small “pilot project”. In other words, start with one master data domain such as the customer data, implement a solution for it, prove to the business that MDM does have benefits for the organisation etc before moving on to expanding the MDM initiative and adding other MDM domains such as products or suppliers etc.

However, whilst implement the pilot project it is vital that all analysis and design is done for all domains that you will eventually want to master data manage. This is important, because part of the pilot project some tool(s) must be purchased, configured etc and limiting the focus on only the immediate domain also limits the benefits gained from the investment.

There are various reasons why it has become necessary to manage master data. This blog post will only explain a couple of those reasons, but I decided to publish more reasons in a future blog post.

To begin with, I will talk about lack of having a single repository storing master data and why it prompts organisations to manage master data. Master data is used across an organisation and usually sits across multiple systems instead of a single, central system. Therefore if there is an error in the master data in one system it will also impact the master data stored in the other systems.

For instance, one department may think that a certain customer has not paid his/her bill and therefore will turn the customer to a debt collection agency. On the other hand, another department such as the Finance department has collected the money from the customer. However, because both departments have inconsistent information about the particular customer the process will continue where the customer may be contacted by the debt collection agency, he may ask for a solicitor to act on his behalf etc eventually decide to leave the company as a service provider (after settling the matter). This is a simple example why inconsistent copies of the same data can have a significant impact on a business.

Had the organisation got a single, central repository of clean, consistent master data then perhaps such mishap could have been avoided easily.

The second reason why you need to manage master data is that an organisation may have acquired or merged with other companies over the years. Each company brings their own systems and data stores which generate identifiers in different ways. Often these identifiers are generated by systems. When a company wants to merge the systems it is not so straightforward on the majority of cases. A simple SQL statement to join by identifiers is not going to work. If you are trying to join equivalent products from the two different companies then the product identifiers may be different. Hence a solid MDM tool is needed to deal with the technology aspect of the merger or acquisition.

For instance, if dealing with customer data the MDM tool must be capable of matching and linking/merging the same customer despite spelling mistakes, use of nicknames etc.

Hence managing master data effectively can have multiple benefits such as the following:

  •  a single bill sent to a customer will save the company money and increase customer satisfaction
  • a single marketing letter will save costs and cause less customer frustration
  • a single data repository will help to get a full picture/value of the customer before contact e.g. a debt collection agency
  • an accurate inventory of product data will save costs on storing the products in the warehouse

About this blog

The purpose of this blog is to help you learn/expand on your current knowledge about the various domains which are part of Data Management. This includes Business Intelligence, Master Data Management, Data Quality, Data Governance, Data Integration etc.


About the author

I am Manjeet Singh Sawhney and work as an Information/Data Management Consultant for Tata Consultancy Services (TCS) in London (UK). Prior to TCS, I have worked for Accenture, Tibco Software and Initiate Systems (now IBM). My areas of expertise are Master Data Management (Customer, Product, Reference), Data Integration and Data Quality. I am using this blog to share my knowledge and hope that you will find it useful. Thanks for visiting.


Advertisement & Sponsorship

If you wish to advertise (banner or text links) on this blog or sponsor any blog posts then contact the author using the 'Contact the Author' form below. Please include some details about your advertisement / sponsorship request and a response will be sent to you within 24 hours.

  • Manjeet Singh Sawhney: Thanks Paul for your comment. Can you please elaborate a bit further on what you mean by "SMASH base [...]
  • Paul Cees: I agree, but the user interface for maintenance with IBM MDM for PIM is not optimal. A SMASH based a [...]
  • Manjeet Singh Sawhney: Hi Girish, Thanks for your feedback. The reason why I believe a repository style data hub is more c [...]
  • Girish Nair: Hello Manjeet, Good article on Hub style. This type of MDM implementation can also be termed as C [...]
  • diablo 3: Everything is really open and quite clear explanation of concerns. was truly data. Your web site is [...]

Twitter Updates

Follow me on Twitter

Contact the Author

Captcha image for Custom Contact Forms plugin. You must type the numbers shown in the image