Why MDM Needs Semantic Technology
Neil Raden
Hired Brains Inc. Santa Fe, NM
Tel. 505-466-2202 Mobile. 805-284-2322 Twitter: neilraden
http://www.linkedin.com/in/neilraden
Semantic Technology Conference San Francisco
The Fundamental Problem with MDM?
Lots of talk about data quality, data governance, but it always starts with same old data modeling and trivial concepts of the “truth”
MDM vendors hail mostly from the data integration, data warehousing disciplines (like me), but claim MDM is
largely focused on operational business processes and change management
We have a lousy track record – we never figured out how to get people to use Business Intelligence!
Accessible versus Usable
-125 bushels/acre
-$8.50/bushel dried at the co-op -$1000/acre (before expenses)
-$4.29/box (18oz) on the shelf -12.9 oz milled corn
-Cost of corn: <$.14
Corn Corn “Flakes”
It isn’t the corn, it’s the box;
Evolution of Data Integration
Data Federation/Virtualization
Not really the answer
What do the arrows do?
What Do They Mean by “Semantics”
(LOL) • Common MDM terms: “semantics” or“semantically consistent”
• But most of their work is in Excel and ErWin
• Semantic (Web) Technology and Ontology
should be part of MDM in at least two places:
• At the beginning, to model and understand data
• At execution time, to reason
• (and it wouldn’t hurt to avail themselves of already
Shortcomings of Current Data Integration Practices
Then you have to jackhammer them up as you go
Starts with a data model
design; only fluid when being poured
Standardizing “Semantics” for MDM*
Standardizing semantics is a process of these steps: 1)Identifying the business process uses of common
business terms
2)Documenting a definition of the term within each business context
3)Determining and documenting those definitions that are equivalent or consistent (and there may be more than one!)
4)Identifying and qualifying those uses of the business term in which the definition is not consistent
Definition vs. Meaning
Definition of a dog:- A domestic carnivorous animal with a long
muzzle, a fur coat, and a long fur-covered tail, whose characteristic call is a bark.
Meaning of a dog:
Definition vs. Meaning
-Neil Armstrong -Apollo 11
-July 20, 1969
-Tranquility Base, Moon, 90210
-First human to step on another planet -End of the “space race”
-Healthcare diagnostics & therapeutics -Microelectronics
-Conspiracy theories: where are the stars?
Definition
MDM Is Too Inward-Looking
Most discussion about MDM concerns:
- Data in operational silos
- Data “governance”
- Data “quality”
- Data “stewards,” “black belts,” etc.
What about data that isn’t in the enterprise? What about “big data?”
Inference: Streaming, External “Big Data” Issues
Example: Pattern vs Semantic
Ducati, 999, Silicone Hose Kit, Blue Blue Silicone Hose kit for Ducati 999 Silicone Hose kit, for 999 Ducati, Blue Hose Kit, Blue, Silicone, Ducati, 999 HseKtBlu-Si, Ducati 999-12/98
expected record
ERROR LOG
Field level matching cannot reconcile
Ducati is a motorcycle; 999 is a model of a Ducati;
Mortorcycles use hose kits; Hoses are made from silicone; Silicone hoses have color; Blue is a color
You can’t always be there to interpret the data. What would your system do with this…
Why Is Context Is So Important
“Katy Perry and Russell Brand
are now officially husband and wife.”
She doesn’t look like a husband… But neither does he, actually.
Meaning, Relations and Reasoning
Summing up:• MDM attempts to tackle the definitional part of
cross-functional data, but rarely gets at the meaning of it
• Meaning is derived through relationships
• E-R models (sort of a misnomer) are not models,
merely representations
• An ontology like OWL is an active model, with
Abstraction, Agility & Inclusion
Loose- Coupling
• Persistent cache
• Temporary cache
• Views, EII, Schema
• DW/DM, MOLAP
• Cached Results
ETL/EII: Directed at conceptual models
Metadata Conceptual-Physical Translator Conceptual Models Reference data Legacy, ERP/ CRM, Web Services, MQ, external Rules Ontology
Prediction Is Part of Decision-Making
“Automated decisions are probably already being used in your industry and they will undoubtedly grow in importance. If your business needs to make quick, accurate decisions, and on an industrial scale, you need to read this book” Thomas Davenport, Professor Babson College
Author of “Competing on Analytics”
“James Taylor and Neil Raden are on to something important in this book – the tremendous value of improving the large number of routine decisions that are made in organizations every day.”
“This blazes a new trail in the crucial territory of finding the business value in systems”
Neil Raden
President & Practice Director
Hired Brains, Inc.
Email: [email protected]
White papers: http://hiredbrains.com LinkedIn: www.linkedin.com/in/neilraden Blog: in transition
(Office) +1 505 466 2202 GMT - 07:00 Mountain Time (Mobile) +1 805 284 2322