Blogs

Are you a master data chef?

Director of Information Integration and Governance Worldwide Sales, IBM

Picture this scenario: a group of chefs at a French restaurant are creating a culinary experience for their dinner guests. First, they select which French dish they are going to prepare. Based on that decision, they decide which ingredients they want to select. Then they cook the ingredients, blend them together, arrange them and present them to deliver the delicious feast their customers are hopefully longing for. In other words, they cook for a purpose—the culinary experience. And quite likely their guests chose to dine at their restaurant for that same outcome; otherwise, they wouldn’t go to a French restaurant.

You must be asking yourself, what is this guy talking about? What does governance and master data management (MDM) have to do with cooking?

http://www.ibmbigdatahub.com/sites/default/files/master-data-blog.jpgThe outcome is what matters

For the same reasons illustrated previously, do not engage in a governance project or a master data management (MDM) project. Look instead for a project that will have an outcome. Examples include risk-mitigation projects, loyalty projects, up-sell or cross-sell initiatives, projects to improve marketing campaign effectiveness or multichannel projects. Governance and MDM projects are meant to solve a complex business problem. For that specific reason, if the outcome isn’t clear, they will very likely be driven by IT. And they will be perceived as complex and painful if they do not bring clear value to the business. Even worse, they may just never happen or fail.

This exercise is not just rhetorical. Many CIOs and CDOs I’ve met with mentioned that lines of business managers sometimes don’t understand the benefits the technology can really do for their business, and we have to spend time helping them run value-assessment workshops. Those same CIO offices that are dying for better governance also mentioned that they struggle to get funds for their initiatives. Face it, very few clients want to do just a governance or an MDM project. To them, this type of project represents a technology term; these projects have no beginning, no end, no proven outcome and ultimately little-to-no budget.

MDM as art

Consider MDM as the art of building an entity—that is, building a golden view of something: people, products, contracts, organizations and the relationship between parties. If you want to succeed, you absolutely need to start by defining the business benefits and quantifying the business values.

I recently met with a client who was interested in building a household view. This project can be pretty complex because a household means different things for different groups in the organization. For a risk-mitigation project, a household may be defined as a group of people being married under a specific contract. For a legal department, a household may refer to a way of identifying parties that are beneficiaries of a given contract, in case the other party dies. Or it can apply to identifying those covering payment in case a party fails to honor a loan payment, for instance.

A household might differ for loyalty management as well. Households can mean all the people living at the same address—grandparents, kids, relatives—even though these individuals might not have a contractual relationship, enabling you to offer products and services to more people living under the same roof. As a result, when you are looking to choose an MDM solution, clearly understanding what are the main business outcomes is crucial.

However, comparing MDM solutions available in the market can be tricky. Marketing teams do a great job at focusing on easy problems to solve and fancy things one can achieve. But several domains exist that you need to pay specific attention to when choosing a tool. Just as chefs carefully pick selected ingredients to prepare a dish, you also need to carefully select your solution’s components. Not all salads are created equally, after all, and they don’t all taste the same.

Solution components and differentiators

You need to pay attention to six key components—and IBM differentiators—of an MDM solution.

Matching engine

You are going to need the best matching engine possible to be able to automatically de-duplicate or reconcile as many records as possible. Think about it. If your matching engine is only good enough and leaves you with 10 percent unmatched records, you may end up with thousands of records to associate and match manually. For a data set based on 10,000 records, that task might be doable, and a team of data stewards can likely do it in a couple days. But no one has 10,000 records. Clients have millions, tens of millions, if not billions of records. For technologies such as fuzzy logic matching, a probabilistic matching engine can prove extremely valuable.

Customization and testing

Provide a data model and out-of-the box, pre-developed services and integration tools that require as little hand coding as possible. Of course, having a data model that reflects exactly a client’s business process and organization is always nice, but who will maintain it over time? Who will pay for necessary technology upgrades, non-regression tests, functional tests, performance tests and all maintenance tasks that have no value to the business.

IBM Data Model provides hundreds of prebuilt services, and it can be customized. It offers the best of both worlds because IBM is in charge of the challenge of upgrading and testing all the components of an MDM architecture. This task is critical because many MDM solutions quickly become the backbone of a client’s IT system for many years. The number of clients that are running old systems because they don’t have funding from their business to do upgrades is quite remarkable.

Cloud computing

You should leverage cloud computing technologies where possible to accelerate your time to outcome. Having a tested MDM solution on the cloud helps deliver quick wins and provide business value faster than ever. As a client, you can choose to run the production wherever you want: on the cloud, on premises or some environments on the cloud and others on premises. This flexibility is a key benefit of the IBM hybrid cloud strategy. We see many projects in which developers work in parallel on several business cases in an agile manner and help reduce overall project time to outcome.

Collaboration and productivity for data stewards

Building an entity and integrating it into a system of engagements is part of a process. However, people need to collaborate to make it work. IBM MDM embeds a powerful business process management (BPM) engine under the cover of the stewardship center. This engine can be customized to fit each client’s process. Many MDM solutions do not cover this requirement very well or are limited, and clients are therefore forced to procure a BPM engine a couple of months after having invested in an MDM solution. Having a collaboration tool among stewards quickly becomes the number one factor for user adoption.

Well-suited total cost of ownership

This component may sound obvious, but the best culinary experience may not be the most expensive one, and for the same reason you don’t want to go to a French restaurant and pay for bread on top of your cheese. Don’t choose an MDM solution that doesn’t provide all supporting programs—databases, process engines, connectors, business rules engines, stewardship consoles and so on—embedded into it.

Organization vision

Make sure you clearly understand the vision of where you want to go in the long term. Some elaborate solutions are good MDM analytics solutions for creating lists, graphs or a single view for an analytical purpose. But they become impossible to manage if you need to integrate the golden view into systems of engagements or systems of records—especially at scale. Also, think about whether you might be interested in context computing, augmenting your single view with external data to find nonobvious relationships or identifying degrees of separation between entities. And consider whether you want to use the entity in conjunction with a complex event-processing engine to create real-time events such as fraud detection, security alerts, advanced analytics and personalized event marketing based on user context.

The feast of outcomes

Now I know what you are probably thinking—what a headache. The good news is, if you opt for the right solution to deliver on your outcomes, the experience can be as enjoyable as fine dining in a French restaurant. We’re open for data and want to make data-driven outcomes easy. Get more information on IBM MDM solutions, and be sure to see the list of 10 ways MDM delivers sustainable competitive advantage.