DM Radio: EDW Mistakes to Avoid

Last week I had the opportunity to participate in a DM Radio interview.  There was a panel of several of us industry types being asked question by the host Eric Kavanagh.  He was assisted this time by co-host Robin Bloor.  The topic was “Mistakes to avoid with an Enterprise Data Warehouse”.  If you missed interview you can follow this link to hear the recording (http://www.information-management.com/dmradio/-10019278-1.html).

Last week I had the opportunity to participate in a DM Radio interview.  There was a panel of several of us industry types being asked question by the host Eric Kavanagh.  He was assisted this time by co-host Robin Bloor.  The topic was “Mistakes to avoid with an Enterprise Data Warehouse”.  If you missed interview you can follow this link to hear the recording (http://www.information-management.com/dmradio/-10019278-1.html).

I bring this up as it was a lot of fun and it was very informative.  As the questions went into several directions I thought I would share with you the other points that I did not have time to get to during the show.  Just a few pointers on mistakes others have made and everyone can learn from rather than make them again.  I have expounded on some of these in earlier blogs but I repeat them  as having in one place is good and you can never hear this enough!

[Disclosure: I work for Teradata, which sponsors the Smart Data Collective.]

1) Mistaking Consolidated with Integrated

By this I mean that people try to make an EDW in name only.  They bring the data marts and consolidate them together on a single platform.  They do not do the hard work of actually integrating the data models and removing redundancy or inconsistency.  Yes, they may be able to get some new functionality and perhaps some marginally better analytics.  However, they are going to be more problems as users start to really be exposed to the data issues that then need more, and harder, efforts to resolve.

2) Mistaking “Phase 1 success” with “Foundational success”

This is what I describe as “short term gain but longer term pain”.  Many companies will have a very good plan to start the EDW but then when push comes to shove they will compromise the data model or eliminate data quality or IT checkpoints in order to get “Phase 1” complete.  Unfortunately they often end up not being able to leverage the phase 1 success into an on-going sustainable, evolutionary, information environment.

3) Mistaking Regulation for Governance

Regular readers will recognize this as it was my last blog: (http://smartdatacollective.com/rob-armstrong/37668/governance-if-it-isn-t-logical-it-s-political).  Just repeating the highlights; governance is when the executive groups, IT management and Business owners agree on how the environment will be prioritized, managed, managed, and utilized.  Regulation is when there is a rule for e very conceivable situation and each new issue generates yet another (sometimes conflicting!) rule.  This leads to an over cumbersome set of processes which then kills innovation and forward movement of the data warehouse.

4) Mistaking Report for Action

The goal of the data warehouse is to provide business users insight into situations so they can take action and either avoid a problem or very quickly identify and resolve a problem.  Too many people start using the system for reporting and then fail to move forward for several of the above reasons.  Reports simply tells what happened to far after the fact to actually impact the business in a serious way.  The data warehouse has to be able support analytics in the heat of the moment from a wide variety of viewpoints, after all that is what the E stands for in EDW.

5) Mistaking Projects with Programs

An enterprise data warehouse needs to be seen as a corporate wide program that deals with the integration, management, and usage of the data.  Unfortunately the EDW is initially treated as another IT project, at the benefit of one group, and it carries the funding and governance aligned to a project focus.  The downfall is that very few projects care about data integration across the enterprise.  As the EDW tries to evolve there are arguments as to which project will fund the integration?  Which project will fund the management and monitoring of enterprise service level agreements?  Which project will run the data roadmap and quality resolution?  Without a proper, longer sighted, program perspective the EDW is doomed from the start.

So, check out the interview to get some other insights from the other panelists.  Also, look around your organization and information implementation.  Are seeing some of the mistakes listed above?  If so, do you have plans to rectify them and get your efforts back on the right track?