Tag Archives: MDM platform

Webinar: Learn How Kodak Is Enabling Customer Centricity With MDM

If you’ve wondered exactly what type of impact master data management (MDM) can have on improving B2B sales here’s a terrific opportunity to find out first hand!

I’m delighted to announce that Kodak, the multibillion-dollar global imaging company, will be joining us for a must-attend webinar on how it’s using a multidomain MDM solution to enable customer centricity and empowering its sales, marketing and customer service teams with access to a single customer view. (more…)

FacebookTwitterLinkedInEmailPrintShare
Posted in Business Impact / Benefits, Business/IT Collaboration, CIO, Customers, Data Governance, Data Quality, Enterprise Data Management, Informatica Events, Master Data Management, Operational Efficiency, Uncategorized | Tagged , , , , , , , , , , , , , , , , , , , , , , , , , , , | Leave a comment

Leading Research Firm’s Take on Multidomain MDM in Its 2010 Predictions

Just this week, leading research firm Gartner, Inc. published its 2010 predictions for MDM. There is one prediction related to multidomain MDM that I found particularly interesting. It mentions that the number of companies shopping for multidomain MDM solutions has increased. Now, why is that?

To get some insight into MDM purchasing and implementation trends, we simply need to look at companies that began their MDM journey in the past five years, especially those companies that started off with a single domain, such as customer data.  Many of these MDM pioneers have since expanded their implementation to other domains such as finished products, materials, price, employees, and so on. But how did they do that? By using the same multidomain MDM platform? Or by separately implementing distinct single-domain MDM applications, such as one for customer data and another for product data?

Gartner contends that no vendor has a comprehensive multidomain MDM technology that handles all different industry use cases using different data domains. A true statement if you are purchasing an “MDM application.” Similar to packaged applications, like ERP or CRM which manage back-office or front-office operations, purpose-built “MDM applications” that focus on a single data domain for a certain industry can, in fact, only handle use cases that are specific to that data domain. So, Gartner is right in saying that a customer that uses “MDM applications” will have to work with different MDM vendors and technologies.

However, this should not be the case if you use an “MDM platform.” We can think about the situation as similar to database or web server technology; these technologies are pretty horizontal and flexible enough to address just about any use case in any industry. In the same way, a multidomain “MDM platform” is flexible enough to accommodate any data domain, and has the ability to cleanse, enrich, match, merge, and display data relationships across multiple domains.

What I don’t agree with in the Gartner predictions is the statement that only large vendors will provide “stronger” multidomain MDM. In my experience, these vendors are largely packaged application vendors, and coming from that heritage, they currently sell different single-domain “MDM applications.” While they talk up “multidomain MDM,” their customer base tells a different story – they have to use multiple distinct MDM applications because no single MDM application can accommodate diverse use cases involving different data domains. In contrast, Siperian has customers in different verticals using our multidomain “MDM platform” to manage multiple data domains on the same platform. Our customers don’t need different “MDM applications” because they’re fully capable of implementing multidomain MDM on their single Siperian platform.

Stay tuned for my forthcoming blog discussing the differences between the “MDM application” and “MDM platform” approaches for cross-industry, multidomain MDM use cases.

In the meantime, what do you think?

FacebookTwitterLinkedInEmailPrintShare
Posted in Master Data Management | Tagged , , , , , , , , , , , , , , , | Leave a comment

Securities Master – Better, Faster, Cheaper.

In my previous post I made the bold statement that I believe that MDM can be used to deliver a better Securities Master solution than the traditional securities master applications, and it can do so faster and cheaper. Thanks for all the feedback and comments, to continue the discussion I am outlining the basic steps of the process.

Create a data model.  In my experience, the best way to do this is to start with a logical model containing an entity (or class, for Java/C++ folks) for all the types of instruments you want to include.  You can make the buckets as big (e.g. “equities”, “debt instruments”, etc) or small (e.g. “common stock” , “preferred stock”, “depositary receipt”, 25 flavors of bonds, etc) as you like — at this point, it doesn’t matter.  Then, take all the attributes from the sources — third-party data feeds, internal data sources, etc — and populate the entities with the attributes. Finally, denormalize the data model so that attributes aren’t duplicated all over the place, and optimize it for the expected query load (MDM systems tend to be read-heavy, so err on the side of denormalizing to improve read performance).  This process — even for large numbers of instrument types (say 400-500) and high numbers of attributes (say 5000-6000) should take a few weeks to perhaps a month at the outside. 

Define the Mappings.  Once the data model is instantiated take data from the sources and run it through transforms, cleansing and standardization to convert everything into the target MDM model.  There are tools available that can speed the mapping process by discovering relationships between the data in a source and the data in the target.  But even when this is done by hand, defining a mapping from a source to a destination entity using a graphical, drag-and-drop transform definition tool should only take an hour or so.  All the mappings for a reasonably-sized implementation (say 10 data sources) would represent about one man-month of work.

Define the match rules.  Matching financial instrument data is either very easy (when the two records have the same IDs, such as CUSIP, ISIN, SEDOL, etc) or hard (when there are no IDs to match on).  When no IDs are present, you need to use other identifying attributes, such as Offer Date, Initial Offer Price, Initial Quantity, Issuer and Name.  Matching financial instrument names requires a very sophisticated, heuristics-based match engine to support “fuzzy” matching handling examples like Options and Commercial Paper, in which the names of different instruments vary only a little — sometimes less than two different names for the same instrument.

Reap the benefits. Since an MDM platform already provides features such as cross-referencing, lineage, matching, merging and un-merging, etc , you can create a securities master in a fraction of the time compared to the monolithic securities master applications.  Furthermore, the end result — a solution that is the right size, tailored to their business, and is flexible enough to change as the business needs change — is far superior. 

FacebookTwitterLinkedInEmailPrintShare
Posted in Master Data Management | Tagged , , , , , , , , | Leave a comment