Seven best practices for primary data management (MDM)

When considering subjects such as primary Data Management (MDM) and Master Data Management, it is the only natural way to seek consultations from people who have already experienced or experienced primary Data Management.

However, the best practices of primary data management are still in its infancy, because it is not easy for experienced organizations to talk about their primary data management experience. Kalido, a master data management technology vendor in Burlington, Massachusetts, also believes that it is very difficult for customers to talk about this topic in the press.

All successful primary data management projects are kept confidential, so it is not easy for enterprises to find the best practices, this is part of the reason why Kalido initiated a customer audit and master data management best practices study by the Ventana Research Institute of San Mateo, Calif. The investigator investigated the best final practices of five anonymous Kalido customers to deduce the conclusion. For the success of primary data management, the Ventana Institute, an experienced consultant and a European telecommunications manufacturer finally provided some best practices and worst practices ).

1. Involve the business or manage it ).

"Primary data management must be driven by enterprise needs, or it may become another database that must be synchronized with all other databases," said David Loshin. He is the president of Silver Spring Knowledge Integrity, Maryland. The company provides primary data management policy development service consulting and enterprise-level policy research.

Similarly, Ventana research found that business personnel rather than IT personnel are driving the primary data management process. Ventana found that it is vital to achieve the success of primary data management from Class C supervisors to senior managers to extensive support for business end users. It is often difficult for us to motivate organizations to move towards the prospect of primary data management, but in the long run, it is very important to start support within the enterprise scope early, users said. If the company's important goal is bound to the project through a stable business case, it should be easy to prove its benefits and boost morale.

2. Adequate time is allowed for evaluation and planning.

Kalido users told Ventana researchers that they would plan at least three months of assessment, talk to customers, and use real company data as examples to build a value verification project. The user said, do not underestimate the time and the process of developing the basic data model.

"This is more complex than people think-and it needs to start early and plan to use real data," said David Waddington, a vice president and Research Director of Ventana, he has been engaged in this study.

Ventana found that IT companies' cooperation is a matter of concern because some companies have experienced project delays due to licensing and access permissions.

3. Set a big goal, but take a small step.

Users told Ventana To consider the basic objectives, but they would limit the vision of strategic deployment. They suggested that once primary data management is easy to use in one place, it should be further expanded. They said that business processes are often mitigation factors rather than technologies, so it is important to get input from end users early in the process.

"If you are only interested in keeping the customer data consistent and sexy, you must oppose it. 'How can I manage all my primary data for a long time? 'The premise is that this is very important, 'Waddington explained. "In the end, you will not have to connect a whole bunch of solutions together ."

4. Potential performance issues

Loshin warned that performance is like a 800-pound orangutan quietly lurking in the discussion of primary data management.

Different architectures can cause different performance losses. For example, if a company uses the master hub Mode for data management, it can become a bottleneck when creating a record that flows through a unique point. Moreover, as applications are increasingly dependent on primary data management, the priorities and running sequence of workflows and systems will become an urgent issue to be considered immediately. Loshin said there are multiple options for companies to solve potential performance problems because they are essentially related to their unique architecture.

5. Establish data management policies and processes.

Experts agreed that time and money should be invested in manpower and process change management, so do not underestimate the workload. According to Roderick Hall, Senior Project Manager, Ericsson, Swedish telecom equipment manufacturer, data management policies may be quite difficult to meet.

A long time before the deployment of SAP primary data management, SDE Moore's company established a primary data group to manage key data assets. IT is a "Shared Service" group that provides services for IT and commerce at the same time. The Group was just established as part of the finance department, but as primary data management received company-wide attention, its functions also changed, Hall said. Their work is not always easy.

Although some departments, such as finance departments, see the value of centralized primary data management, their groups are reluctant to give up their rights to manage data.

"People have given up their freedom to manage their own data and put their own primary data to the primary data group of sdegemel, Which is local. To get people's recognition of this fact, they really had a tough fight, "Hall said.

6. carefully plan the deployment.

Ventana found that primary data management is still relatively new, so training for business and technical personnel is more important than ever. Waddington says that the use of system integrators who are not trained or are quasi-trained and

Outsourcing attempts may cause major problems or even project delays for primary data management users.

7. Consider the transition plan.

With the prospect of a project being rolled out, the project will have some impact on the key processes and systems-No worries for no reason. Loshin recommendation companies should plan a transition policy for primary data management to allow static and dynamic data synchronization.

"Trying to adjust the potential infrastructure without affecting daily operations may be as challenging as making holes on the road without interrupting traffic ". Loshin said.