If we take the key goal as being the creation a consolidated accurate view of enterprise data and using that to drive enhanced business decision-making, leading to improved business performance and increased business agility, then some projects are wildly successful.
Others are yet to prove business value. While the potential benefits of MDM are clear, actual success depends on a number of factors.
Let’s consider five key reasons why MDM projects can go wrong and briefly look at how a new model for master data can deliver value for customers.
- Leveraging other IT and business systems
The technology within MDM is a sometimes a mish-mash of several technologies that have come together over the years. MDM is about accessing data, creating a central repository and matching records and identifying a golden record or single version of the truth, among other things.
However, there are also related technologies like business process management (BPM) are used in the process. Companies are increasingly trying to understand how MDM can be integrated to deliver added value in Big Data projects and vice versa.
Companies even find success in MDM with technologies like web portals and wikis to communicate the status data management. It’s a complex co-mingling of technologies, combined with business processes that make MDM successful.
However, it is not a given that vendor selection committees look at the big picture with regard to their future with MDM. Features are so important that companies will spend time to produce RFPs with extensive feature lists.
However, MDM customers should carefully look at the ability of the architecture to support non-MDM technologies and fringe technologies that support MDM. These systems must come together in order to succeed.
Companies think beyond the standard set of features needed in MDM. They should consider both how the vendors’ technology comes together, since many commercial MDM solutions are mash-ups themselves, and how it can ‘play nice’ with other technologies.
- Spending too much or too little time on data governance
The success of MDM is often closely linked with an organisation’s data governance strategy, in which organisations put in place formal processes and responsibilities for establishing the management of data. All businesses can benefit from having a proper data governance approach in place.
However, it is particularly critical for MDM projects. Data governance helps align cross-functional teams and key stakeholders to support MDM. Without this alignment, it is difficult to achieve success with MDM in most cases.
Even if organisations have successfully linked all of their business systems together, if business users can’t find value, the final output of the MDM strategy will be meaningless.
While plentiful resources can be a success factor, there must also be restraint on the scope of the project. MDM can suffer from scope-creep, where the project goals are constantly expanding, or big-bang, where the scope is set too high.
On the other hand, they can also be too small to prove valuable to the organisation. Make sure you work with your business users on an MDM strategy that can provide provable business value in a relatively short amount of time.
- High costs, limited sponsorship
MDM has traditionally been expensive in both licence cost and resource allocation. The typical service-to-licence ratio has been 3:1 and even 9:1 for some solutions, meaning the cost of consultants can be more significant than licences.
After all MDM is often a complex process and the solutions that support them often require costly levels of customisation and integration Combine this expense with the average sales price for an MDM licence €’ anywhere from $500K to 1.2M €’ and you are looking at a very expensive project.
Open source solutions are changing the service-to-licence equation. Much of this comes down to the innate flexibility of open source. With an open source solution, such as Talend MDM, users have access to the source code.
The solution implements an open API and they can develop simple extensions to more easily modify or customise the solution to suit their needs. You still will need consulting, but many open source users find this cost to be even lower than commercial software due to the ability to customise.
- Single track focus
Most MDM strategies include a series of coordinated projects that focus on the customer and product areas where ROI is greatest and therefore more easily justified. The plan may be to start with a single data domain, or even a few attributes, and improve value. When you’re done, you can move on to the next domain or attribute.
However, keep in mind the long term costs of this strategy. You don’t want to have significant incremental costs as you expand from say, customer data to supply chain data. Legacy technologies tend to focus on one domain or another.
The new breed of MDM solutions addresses this issue by adopting both a flexible data model and a flexible approach to pricing solutions to bring down the cost barriers and enable ROI. Business can better benefit from your MDM strategy, not only today, but also tomorrow.
- Agile Approach to MDM
Finally, MDM strategies often touch lots of different areas and as such there can be ownership issues which affect progress. Businesses will find achieving success easier if they focus on picking high-value; low-cost projects (small proof-of-concepts rather than major business-wide initiatives) to begin with, start proving the value of those projects and then leveraging the ‘little wins’ they make into other projects and domains.
You may find success in identifying these projects by being social and connecting with your colleagues. Bring up the issues that they may be having with data management and be willing to help with them.
Track costs and ROI and make sure you market the value you’re bringing to the organisation. It can only help in future corporate sponsorship.
A New Model for Master Data
So how can organisations get to value more quickly, with less pain and money? The latest breed of open source MDM solutions address these integration, data governance, time to value and organisational issues directly and effectively.
To meet the evolving requirements in today’s highly competitive business environment, today’s MDM solutions address all these challenges and make the potential benefits accessible to a wider audience.
MDM solutions need to deliver a range of functions from data profiling to data integration and from data modelling to data cleansing. They must ‘play nice’ with other technologies like web portals, wikis and more.
When choosing an MDM solution, it’s important to realise that these disparate components may or may not interoperate smoothly despite being licenced or packaged together. The best MDM solutions compile these functions and integrate them together so they operate as a single solution. This reduces licence cost and lowers the need for expensive consultants.
In this respect, offering a single solution unified approach to data management has significant advantages. One product provides functionality across data profiling, integration, quality, matching, survivorship, stewardship, validation and workflow. This is not only a huge cost saving but a time saving as well, because you negotiate one instead of multiple licences.
Enterprise software typically involves black box, proprietary solutions with functional extensions and customisations to fit your explicit requirements at a given point in time. In contrast, open source MDM solutions give businesses access to the source code, enabling them to modify or customise the solution to suit their needs.
With the lower prices associated with open source businesses can start projects immediately without a long procurement cycle. This approach allows them to realise value quickly and implement base functionality that they can build out over time.
Ultimately, this new breed of solution addresses the key challenges of MDM, by delivering easy-to-deploy, cost effective, rapid end-to-end solutions for master data management. Today, this new model is helping not only to redefine MDM but also to define the next generation of data management projects.
Posted by by Steve Sarsfield, Product Marketing Manager, Data Management Talend