Driving Business Value at Intel
Intel’s director of master data engineering explains the challenges of profitable growth and how MDM can help
How to move to data-driven decisions and rely less on human judgment -- and the role MDM plays in this transition
- By Linda L. Briggs
- March 17, 2009
Intel faces the constant challenge of sustaining profitable growth with products that often become obsolete at a faster rate than they take to develop. Rapidly changing industry conditions also drive complex market segmentation and price pressures. These realities, in turn, demand improvements in information solutions for managing supply chain and customer relationships, all of which are critically dependent on the availability of high quality master data.
In this interview, Greg Valdez, Intel’s director of master data engineering, discusses the challenges of profitable growth and how MDM can help. “We want data-driven decisions with less dependency on human judgement,” Valdez says -- a key factor in achieving a very low-cost supply chain.
Valdez will be speaking on this topic at TDWI’s upcoming MDM Insight conference in March.
TDWI: Can you describe some of the challenges that led you to implement an MDM initiative at Intel?
Greg Valdez: It starts with driving profitable growth at Intel, and using master data management capabilities as part of that strategy.
For Intel, profitable growth comes from success in achieving our business objectives in specific areas, such as the speed and quality of our new product development, achieving a low-cost supply chain, including an effective sales and operations planning process, and driving consumption through sales.
New business models at Intel, such as the Digital Health Initiative and sales force efficiency and effectiveness, [are examples of where we] have to be more effective and innovative in getting design wins and driving our architecture to a broader customer base. It’s those business objectives that motivate everything we do.
So MDM capabilities are playing a key role in accomplishing those objectives?
Yes, MDM capabilities are part of the strategy to achieve those objectives.
There are compelling value propositions associated with MDM, notably better visibility -- enabling broader, more-complete views of the business. That cannot be achieved without semantic consistency, which enables business activities and decisions to be linked and integrated.
Most decisions and activities have a substantial information content, which can be traced back to granular data. At Intel, we are always looking for ways to execute better and eliminate workforce inefficiencies associated with poor information quality and inconsistency.
We get better agility by enabling greater flexibility through separation of core reference data from its usage. There’s an emphasis on that basic premise when we create a shared service for data -- and MDM is the platform we are using to deliver those shared data services.
One interesting thing about your MDM implementation is that it’s very real-world -- you’re focused on profitability.
Very much so. I talked earlier about profitable growth coming from success in achieving business objectives in specific areas. The second reason that we’ve put such a focus on business information value falls under the general heading of business intelligence competence. We are working to grow that competence and apply it to things like global indicators and supply chain management.
One of our biggest challenges revolves around our [recently announced] Atom Processor. It’s a remarkably designed and attractively priced processor that enables us to competitively pursue segments such as “netbooks” and mobile Internet devices. From a supply chain standpoint, it’s a fundamental game-changer. We are working to ramp it to very high volumes and part of that plan is to create new business intelligence capabilities that drive costs down and improve execution. There is further opportunity to extend these breakthroughs to our higher-end products as well.
Was there a particular place within Intel that you started with your MDM project? Was there an area where you said, “Let’s focus here first”?
We picked a handful of subject areas that relate primarily to the supply chain. These include items (what we make, buy, and sell), suppliers, locations, and commodities. This is where we have the strongest pull for higher quality data due to its alignment with business priorities.
We have put less emphasis on customer data up to this point but are now shifting our focus to that area. It will present several new challenges and will require a tailored solution that won’t be an exact copy from the other subject areas.
Where are you in the process now?
We have MDM solutions in production for four subject areas that fully reflect our enterprise data standards. We have used a variety of designs and technologies in doing so; it hasn’t been a one-size-fits-all approach.
For item, supplier, and commodity subject areas, the MDM solutions employ the use of work flow and business-rules-driven transactions for the create, update, and delete processes. They are full-blown applications, and not just consolidation and cleansing capabilities. We are seeing excellent results in terms of usability, business productivity gains, higher data quality, and easier integration. Since we went live we have processed tens of thousands of master data transactions and have a business user base in excess of 200 (and growing).
So you have come quite a ways.
We have come a very long way. What started several years ago as a data quality initiative has matured nicely into a business capability that really delivers value. We’ve only begun to see the benefits of making decisions based on a common data set, but the prospects are tremendous.
One remarkable thing about this project is its size. How long did all of this take?
It’s been a long time in the making. As a running program, it has taken us about two-and-a-half years. We will continue to evolve our capabilities and drive improvements based on actual usage. One of the benefits of our designs is their inherent flexibility. I’m anxious to see how well we do in responding to business changes.
I wanted to go back to something you said earlier about having plenty of business support for your efforts with MDM. How key was communication in the rollout in strategy?
We have spent a lot of time on communication -- not necessarily to increase the volume of communication, although clearly you have to get critical mass, but really to get the right people on board. We focused on creating a really solid “teachable point of view” to do this. That teachable point of view for us has revolves around the cycle of standardize, consolidate, optimize, and utilize. Using the teachable point of view, or TPOV as we call it at Intel, has helped people quickly grasp what we are trying. Relating our road maps and deliverables to these steps makes them more logical and easier to relate to. I like to think of our approach to communication has highly entrepreneurial. You have ten minutes or less to convince someone that your idea has value, so you have to use key words and imagery to get them on board.
What else have you used to explain MDM and generate support?
We’re also starting to talk about the concept of “master data on-demand.” This service-on-demand concept has been used to rally the business and IT to create and use data in a shared-services environment. That really enables the business solution teams to leverage the single-enterprise schema, which can accelerate their time to solution and conserve resources in doing so.
The data services are driven by business priorities and economies of scale. We focus their value on elements of content, quality, and ease of use.
The messaging here is key because one of the things about MDM is that it’s very abstract, very obtuse to many people -- but if you put it in terms such as “data on demand” or “shared data service,” it tends to resonate better. We’ve also leveraged social media and even created some videos that have a nice blend of fun and advocacy.
We’ve tried to target certain groups of people with our communications in a way that will stick. We know it’s working when people start to pick up the terms and use them in their own conversations. When I have a business guy out there saying, “We are almost done with the ROO [record of origin] for supplier data,” I know we have hooked them.