Reports of the Enterprise Data Warehouse’s Death Are Premature

by   |   June 19, 2013 6:23 pm   |   2 Comments

Dr. Bjarne Berg

Dr. Bjarne Berg

The enterprise data warehouse isn’t dead just yet.

Bjarne Berg, co-founder of the business intelligence consultancy Comerit, said, 99 percent of Fortune 500 companies are still using some kind of enterprise data warehouse (EDW). That places EDWs firmly in the land of the living, at least for now.

Berg  is an SAP University Alliance professor, and said he helped implement a significant number of those warehouses in the first place as an SAP partner.  Now, he said, many large companies are starting to move away from the rigid data modeling of the traditional EDW and toward faster and more flexible technologies like Hadoop or in-memory databases.

But just because results come faster and there is significantly less data modeling involved in the newer approaches doesn’t mean some core concepts if the EDW aren’t still crucial, like data integration and cleansing, historical reporting and security. It’s just that business users need their reports more quickly today than the monolithic, schema-based warehouses can provide.

In this interview with Data Informed staff writer Ian B. Murphy, Berg discusses how the EDW as we know it is changing. He also describes what a new enterprise-level data store looks like today, and how the drive for self-service business intelligence, and increases in database performance, have made the traditional EDW model more cumbersome than useful.

Email Staff Writer Ian B. Murphy at Check out other Data Warehouse podcasts from Data Informed.

Related articles on Data Informed:

Alternative to in-memory analytics relies on column-based data stores. 

Considerations for storage, appliances and NoSQL systems for big data analytics management.

League of Legends powers up Hadoop-based queries with Platfora.

Tags: , ,


  1. EB Quinn
    Posted June 20, 2013 at 12:03 pm | Permalink

    Could not agree more with Ian on this. Even organizations that throw in with big data/advanced analytics will need their warehouses as a primary data source. And plenty of those data warehouses now have connections to Hadoop and MPP Analytics solutions too (e.g. Teradata). And existing DWs often reflect the process and data governance work the organization has already performed, it would silly to move all of that good work to the recycle bin.

  2. Josh Andrews
    Posted April 3, 2014 at 5:13 pm | Permalink

    I’ve just finished a large data warehousing project with a huge non-profit. Perhaps 5 or 10 percent of the problem had to do with inflexibility of schemas and ETL software, etc. 90 to 95% of it was dealing with data cleansing problems, requirements definition, user education, change control, etc. This “death of the DW” fad is the same kind of thinking that leads people to believe that the reason that programming is hard is because the languages are just too arcane and if you could make a better programming environment then “normal people” could code. Not true — most of the difficulty is in the clarity of thinking and the concentration, not the programming language.

Post a Comment

Your email is never published nor shared. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>