Data Warehouse Survey Update

It’s the end of October now which means that the autumn trade show season is drawing to a close. This year we sponsored and exhibited at the TDWI World Conference in San Diego (and in Chicago earlier in the year), and the Teradata PARTNERS Conference in Nashville. Once again we gathered survey responses in our booth, so I took a look at how things have changed, or not, since last year’s update.

At the Teradata event we saw more people using the Teradata logical data model (LDM) as a blueprint for their warehouse – 59% vs. 41%, compared to only 49% using the LDM as reported in 2013. I find this encouraging, because according to Teradata, these models are designed to “eliminate the stovepipe approach to accessing data and, more importantly, gets your IT team and business users focused on connecting their individual pieces to a broader company strategy.” The challenge of course, is HOW to implement a logical model, as it is only a blueprint – someone has to bring the tools to actually build it.

TDPUG survey on resources


We also asked about maintenance costs of the Teradata warehouse, and the results were once again on the high end of the spectrum, as compared to results we saw from the more general TDWI conferences. 40% reported spending over $1 million annually, with the balance spread fairly evenly across the other cost ranges as reported in 2013. But interestingly, I noticed that for individuals reporting they use the LDMs, the vast majority reported spending in the over $1 million range – and they also tended to report requiring 20+ people to support the data warehouse environment. This leads me to conclude that despite the logical model as a starting point, it doesn’t save resources. Or put another way, the resource requirements stay about the same.

Additionally, having a blueprint didn’t necessarily make it faster to integrate new data sources or handle a re-organization and deliver new view and reports. 90% of respondents at the Teradata event reported it took more than 1 month to integrate a new data source – 48% said it took between 1 and 3 months, and 42% said it took more than 3 months. Similarly, 84% required the same amount of time to deliver new reports and views due to a re-org – 39% required 1-3 months, 45% need more than 3 months. This doesn’t say much about agility or flexibility and appears to indicate that there is still a long way to go to ensure the data warehouse keeps pace with business change.

So what’s the answer? From our point of view, having a model that can actually generate the warehouse design and build will save resources and time to deploy, and maintain. Additionally, using a model that remains “connected” to the warehouse will enable you to make changes such as integrating new data sources and handling re-organizations so you can more easily keep pace with business change.

TDWI now offers a Data Warehouse Automation class where you can learn more about tools that do this. But if you want to know more now, check out our video tutorials on Kalido’s business-model-driven data warehouse automation.

0 replies

Leave a Reply

Want to join the discussion?
Feel free to contribute!

Leave a Reply