This week, I attended David Butler’s presentation at the Oracle Applications Users Group COLLABORATE 11 conference in Orlando, FL. David Butler has been with Oracle for almost 14 years, and has been with the Applications side since 2001.
If you don’t fit MDM into an architectural vision, you’re just throwing “stuff” at a problem.
The data is fragmented – it’s all over the place, in CRM, in ERP, in different divisions, locations, processes, formats, and silos. And it’s in a state of flux, with potentially thousands of points of entry. Master data changes at a rate of 2% per month – across companies, individuals and products. This represents a huge cost of doing business.
But don’t just throw technologies at the problem. There’s an alphabet soup of different tools and products you can purchase, but you can wind up making things much more complex. You want to make things simpler (not more complex).
Start with the applications, the CRM, supply chain, operations, web sites, financials, distribution, ERP, product lifecycle management, etc. People have been writing code for “point to point integration” for many years – the “N squared” problem – the “hairball”. This isn’t sustainable. This is what drove the creation of Enterprise Application Integration (EAI) originally. This metadata layer, with adapters for each application, supports dozens of applications. But the data in the applications, is still inconsistent in format, and this approach still doesn’t address the quality of the data.
More recently, we’ve had the concept of orchestration across the enterprises, using web services and standards across all of these various types of applications. But again, we’re just inventing new ways to break processes. And even more recently, we’re moving things to the cloud, which in some ways, opens up Pandora’s Box, because it moves some things out of our reach into the cloud.
The data warehousing approach moves everything from the applications into an enterprise data warehouse. This has been around for a long time, and is still very powerful. But in some important ways, it doesn’t deliver the “Single View of the Customer”, because it can be a case of “Garbage In, Garbage Out”.
What you need is a Master Data Management solution, which gives you a chance to consolidate the data into the MDM hub, to cleanse it and govern it, and to move it back to the source applications. Then the business processes can use the cleaned and improved local data. You get a chance to improve your business processes. The transactional data finally gets cleaned up, and you can feed that clean data to the data warehouse, and you can do meaningful analytics in a business intelligence solution on top of the data warehouse.
Some companies started building the improved data in an MDM hub strictly for analytical purposes, and feeding it into their data warehouse for BI. But when they saw how good the data was, they started feeding that data back into their operational system too.
The data hub is the data model, plus the purpose-built applications like Oracle Customer Hub, Oracle Product Hub, Oracle Site Hub, etc. (state of the art applications, which Gartner has recognized as belonging in the leaders’ quadrant of the Gartner Magic Quadrant).
Oracle can use its own ETL tool (Oracle Data Integrator, ODI) but can also use other ETL tools as well.
Oracle has a state of the art product for hierarchy management, Data Relationship Management (DRM), which can manage complex hierarchies for many different domains of master data.
Oracle also provides BI Publisher as a business intelligence tool that runs against the MDM hub, to provide useful reporting against the master data that accumulates there.
The EAI that Oracle provides is Fusion Middleware, which provides an Enterprise Service Bus.
With the Sun acquisition, Oracle can provide an integrated hardware stack to run the MDM hub and the data warehouse on the Exadata and Real Application Clustering environment.
Oracle’s recommended Master Data Management process is (a) Consolidate, (b) Cleanse, (c) Govern, (d) Share.
This process allows you to add attributes to the underlying data model, and the data quality tools that are built into the MDM hub will operate against those extended attributes.
Oracle’s suite includes:
- Oracle Customer Hub
- Oracle Supplier Hub
- Oracle Product Hub
- Oracle Site Hub
- Oracle Hyperion Data Relationship Management
Today, the applications have their own individual user interfaces for data governance, but the roadmap eventual will move them to a common data governance user interface.
Oracle MDM high level architecture is based on an MDM framework, with the MDM applications on top, and verticalizations for key industries on top of that.
MDM and Service-Oriented Architecture: MDM without SOA will be a single version of the truth in its own silo. SOA without MDM is an enterprise-wide service without reliable data that breaks at the application boundary.
Forrester referred to MDM and SOA as the most important business case with the most return.
MDM and Business Intelligence: this gives you the quality governed dimensions, the corporate cross reference and the hierarchies needed for accurate roll-ups.
David Butler gave a great example of a consumer getting married and changing her name, two manufacturers who are both subsidiaries of the same D&B global ultimate, two suppliers who are the same but with different D&B trade style names, and two products with different descriptions that are actually the same.
So the example illustrates how a data warehouse can return wrong answers unless you “work the data” through Master Data Management and data governance.
The strategic nature of MDM is:
- Foundation for Customer Relationship Management (CRM), Enterprise Resource Planning (ERP) and Supply Chain Planning (SCP) deployments
- Foundation for Service Oriented Architecture (SOA) deployments
- Foundation for Business Intelligence (BI) deployments
- Key to corporate data governance
- Key to business process optimization (BPO)
- Operationalizes the data warehouses
- Path to IT agility
Oracle has 1,250+ customers, is growing every month, and will have a dedicated MDM track at Oracle OpenWorld this October.
It’s always great to see my friends from Oracle and the Oracle customer and partner community at the COLLABORATE conference, and I thought David did a great job outlining what for a lot of people is a key question – exactly why is MDM such a strategic foundation for a lot of the technology activity going on in larger companies?