Quotes

Quotes about the Data Vault, as well as a quote about the certification class that I teach.

If you have a quote you would like to add to this list, either about my class, or about the Data Vault – please use the Contact Us form, we’d be happy to add it!

If you are interested in Certification, please contact me directlyI no longer support nor endorse any other certification class on the market today.  I have updated the materials, and brought to market Data Vault 2.0 with new ideas, new standards, and new rules.

__________________________

“Dan has devised one of the only truly groundbreaking innovations in information architecture over the past twenty years. The data vault should be considered as a potential standard for RDBMS-based analytic data management by organizations looking to achieve a high degree of flexibility, performance and openness.” – Doug Laney, Deloitte Analytics Institute

__________________________

“A central, data vault modeled EDW gives us the flexibility, expandability and manageability we have lost with our existing BI architecture (data marts on top of source systems). Furthermore, it provides us with auditability and readiness for real-time data which we see coming our way in the future. We are confident our BI architecture will be future-proof now.” Johannes v/d Bosch, (internal) BI Consultant at Waterschap De Dommel (local Dutch government)

__________________________

“The architectural struggle between Inmon and Kimball disappeared when I read the first article about the Data Vault architecture in 2001. From that moment I was convinced that a solid and robust Data Warehouse implementation should be created by a central normalized information factory as the central storage area with a separate dimensional environment, specific for reporting and analysis goals. And Data Vault is the best suitable solution for the implementation of the normalized information factory. It has proven its value within a number of implementations I was involved. It also proved its value, because of the several initiatives to create a packaged solution for a generated data warehouse. Data Vault appeared to be the fundamental of these initiatives.” Maarten KetelaarsManager – Business Intelligence & Data Management – Custom & Emerging, Accenture Technology Solutions

__________________________

“The data vault architecture has enabled us to deliver results sooner by using an incremental, consistent approach to building enterprise data warehouses. The repeatability and simplicity of the loading architecture has reduced the time and cost to develop our ETL.” Bruce McCartney

__________________________

The Data Vault is the most important innovation in data warehousing since the legendary Inmon-Kimball debate.” Tom Bruer

__________________________

I think the Data Vault is the best thing that happened to us after Inmon’s CIF and Kimball’s Dimensional Modeling. It is the first modeling technique that was designed specifically for the Enterprise Data Warehouse.   Don’t be fooled by the methodology however, at first it appears very easy: just hubs, links and satellites. Data Vault is much more than ‘just’ a modeling technique. The real value comes from the Data Vault architecture, supported by the modeling technique. The keywords for me are: traceability, auditability, consistency, historical correctness, repeatability and the possibility for automatic generation of models and ETL-packages. Kasper de Graaf

__________________________

“Having adopted the data vault (DV) modeling methodology early in our BI environment implementation, we have found that there are numerous points of value this approach brings to our business. Perhaps not surprisingly, the business benefit is indirect but very real.

The reality for our BI efforts are that we have to roll out reporting and analytic structures with very limited resources (both infrastructural & manpower). Deciding to take a federated approach and create an Inmon style Enterprise Data Warehouse (EDW) as a staging area from which we would populate a single Kimball style (bus-architecture/conformed dimensions) Enterprise Data Mart (EDM) for our data, we had to accept that this approach requires more effort up front. It wasn’t an easy sell. In our fast paced business environment, short delivery time frames are highly valued. Elegant architecture is secondary to quick project churn.

Accordingly, our first deliverables were more of a repository of denormalized files rather than a constelation of star schemas, since the “flat” tables better fit immediate business needs. We are now moving towards integrating our data into our recently acquired Siebel Data Warehouse (which was thankfully designed extremely similar to our EDM concept). Doing so is going to be greatly simplified by using the already cleansed and organized EDW as a source.

The DV allows us to have a superior BI architecture and simplifies the integrating of data residing in what I lovingly refer to as our “abnormalized” data sources. Benefits for the data modelers include:

  • The DV allows new sources to be quickly integrated into the EDW iteratively as each new project comes through, without dorking up exiting EDW structures. This is probably it’s strongest point.
  • The time variance approach of data organization is exceptionally useful.
    (Normalization typically establishes relationships based on logical associations between the data elements, but does not consider data change characteristics like the DV does. I thus disagree with the assumption that the DV is a “normalized” structure at all. In my opinion it should be viewed as totally distinct from either ER (1st-5th Normal Forms) or Dimensional modeling techniques, even though it bears some similarities to both.)
  • ETL work can potentially be divided between the source-to-cleansed-EDW portion and the EDW-to-data-mart portion of work.
  • The relationship between a DV hub and satellite structure and a dimension table is obvious. Beyond dimensional modeling, multiple denormalized reporting tables with varying grain can be spawned from subsets of a single hub/satellite structure.
  • DV structures are very flexible and reusable. Someone likened it to working with tinker toys, which is a pretty good analogy.

While we do not have an ODS, I can see how data-change-sensitive DV modeling could possibly fit well in that area, especially if it’s a near-real time ODS. History, easily captured via the DV, is typically not maintained in an ODS, so I’m not sure how one would work through that. Might require altering the methodology somewhat… (Has anyone out there tried this?)

Having invested the time up front into creating our DV-modeled EDW iteratively with each new request, we have enabled ourselves to meet new business requirements faster and more accurrately through reuse of EDW structures. Again, the business user needing a report or dashboard metric doesn’t see DV modeling as a direct benefit. It reminds me of the commercial of the company that claims “We don’t make a lot of your products. We make a lot of your products better.” Business users are able to get new reporting and analytics needs filled faster. Being able to integrate our data into Siebel’s DW is going to be easier and will (hopefully) roll out quicker. And in today’s business environment, information accuracy and project delivery speed count.” Kevyn Schneider, Administaff (2004)

__________________________

“As in any data modeling approach i had to query the source data extensively to determine if the vault i was building could break. as i was doing this exercise and extracting the business components from the source data, i slowly uncovered the problems that existed in the source model. the requirement of a hub standing on its own showed me that there were certain entity tables in the source which were really links. i could also begin to see where the source had been scabbed onto to solve certain modeling needs. it became very obvious that the source model was in terrible shape and that the act itself of bringing data into the vault was creating an unbreakable model.
i guess what i’m trying to say is if you are able to load the vault model that you build, then it’s almost certainly modeled correctly. it’s kind of a self fulfilling prophecy. ” Ben Isenhour

__________________________

“Within the Friesland bank, a medium sized Dutch full service bank, The Data Vault is implemented as the foundation of the corporate data warehouse. All datamarts are fully loaded from the Data Vault. The Data Vault serves as the single source of truth. The platform is Unix and Oracle.” The implementation is done by Capgemini BI consultant Michael Doves

__________________________

“This should be called the “Foundational Warehouse Model”, and it looks to be a solid implementation paradigm that’s highly scalable.”  Clive Finkelstein, October 2003 (Works with John Zachman on the Zachman Framework)

__________________________

“The Data Vault is functionally strong and a viable architecture in implementing your Enterprise Data Warehouse” Bill Inmon in 2002

__________________________

“The Data Vault is the optimal choice for modeling the EDW in the DW 2.0 framework.”, Bill Inmon, Father of Data Warehousing

__________________________

“The Data Vault is foundationally strong and exceptionally scalable architecture.” Stephen Brobst, CTO Teradata

__________________________

“In planning to build a CIF infrastructure for Denver Public Schools, we have determined that the Data Vault approach is what we will use to build the central Enterprise Data Warehouse. We believe that this data modeling technique is the best suited for designing a central, historic data repository because of its flexibility to easily add new subject areas and attributes. This will allow us to grow the EDW in an organic manner, over time, as we discover new requirements. ”  Kent Graziano

__________________________

“At Cendant Timeshare Resort Group we were able to reduce our daily volume growth by 85% by applying the Data Vault technique and removing redundant data. In addition to disk space savings, the Data Vault allows us to easily add new information and rapidly create reporting subject areas as business needs change.  The Data Vault modeling approach should be the foundational core of all leading edge Data Warehouses.” Mike Bush, Cendant TRG Data Warehouse Architect

__________________________

“The Data Vault is a technique which some industry experts have predicted may spark a revolution as the next big thing in data modeling for enterprise warehousing…” Doug Laney Nov 4-7, 2002

__________________________

“Due to the nature of the Data Vault architecture, its’ definition of separation of data by semantic grain, type of data, and rate of change – as well as its’ combined effort of integrated satellites and business keys, it appears to be the only data modeling architecture to contain metadata at the cell level. Where every element holds meaning. ” Dave Wells, Director of Education, TDWI (2004)

__________________________

“I just completed the certification class last week. So I thought I’d write a short post about it.

First of all, I’d like to thank Dan Linstedt for offering the class. I found i fascinating, insightful, instructive, and exciting (It’s exciting to be a part of something that’s going to really make a difference for my company).  I would recommend the class to anyone who’s exploring the Data Vault as a potential Data Warehouse environment model. The content of the course was clear, even while it was heavy and a lot of material was covered. The certification exam was difficult and demanding, but then you know that when you pass, you’ve truly accomplished something.

Dan is a gifted instructor as well as a highly knowledgeable and creative data solutions architect. On top of that, he’s a nice guy. I promise that you’ll enjoy the class. I sure did.” David Martin, with regards to my Certification Class

__________________________