Common Enterprise wide Data Governance Issues – #14. No Enterprise wide Data Model

I was reading David Loshin’s excellent post How Do You Know What Data is Master Data? and I thought “I know – I’ve covered that question in my blog” – but I hadn’t.  So here it is.

Your “Enterprise Wide Data Model” tells you what data is Master Data.

Unfortunately, most organisations lack an Enterprise Wide Data Model. Worse still, there is often little appreciation among senior management of the need for an Enterprise wide Data Model.

Impact:
The absence of a Enterprise wide Data Model makes it difficult for even technical experts to locate data.  The data model would distinguish between Master data and replicas, and would clarify whether the data in the model is currently in place, or planned for.  Without an Enterprise Wide Data Model, data dependent projects (e.g. BASEL II, Anti Money Laundering, Solvency II) must locate data (especially Master Data) from first principles, and face the risk of not finding the data, or identifying inappropriate sources.   New projects dependent on existing data take longer than necessary to complete, and face serious risk of failure.

Solution:
The CIO should define and implement the following Data policy:

An Enterprise wide Data Model will be developed covering critical Enterprise wide data, in accordance with industry best practice.

Time to sing from the same hymn sheet

One notable exception to the norm:
This is not a plug for IBM…. merely an observation based on my experience.

I worked in an IBM development lab in Dublin during the 90′s. At that time IBM developed a “Financial Services Data Model” (FSDM). Dublin was IBM’s “FSDM centre of excellence”. BASEL II turned FSDM into an “Overnight success”- TEN YEARS after it was developed. Organisations that had adopted IBM’s FSDM found it relatively easy to locate the data required by their BASEL II compliance programme.

I forsee a future in which all financial services organisations will use the same data model, including Financial Regulator(s).  “Singing from the same hymn sheet” will make communication far simpler, and less open to misinterpretation.

The lack of an Enterprise Wide Data Model is just one of the many data governance issues that affect organisations today.  Assess the status of this issue in your Enterprise by clicking here:  Data Governance Issue Assessment Process

Does your organisation have an “Enterprise wide Data Model” – if so, how did you achieve it?  Did you build it from scratch, or start with a vendor supplied model? Please share your experience.


About these ads

9 Responses to Common Enterprise wide Data Governance Issues – #14. No Enterprise wide Data Model

  1. Henrik Liliendahl Sørensen says:

    Good post Ken. I am actually also at the moment involved in a Data Management initiative where we work on a common data model. This is within public transportation where we have great success in using an industry conceptual data model called Transmodel. Every industry has its own master data with its own labels. In public transportation we have persons in roles as passengers and drivers, we have vehicles, timetables, stop points, services and so on, all reflections of the real world.

  2. Dylan Jones says:

    I think this is one of the most underutilised activities as it provides benefits across so many different IT and business initiatives.

    I use them a great deal in data migration projects as we introduce a common model at the starting point to enable discussion. Particularly when the targets model is poorly defined it really helps us to move forward.

    There are some good examples for Telco on this site:

    http://www.tmforum.org

  3. Rayk says:

    Hi Ken,
    good and clear posting, than you for that. I wonder where you all have your metaphors from. I like that skill. What I also like is the mentioning of IBM’s FSDM, that speaking the same language within the company, but also outside (thinking regulatory institutions). Currently I’m in the position, that the business is promoting such an enterprise model, but IT is not. Good, that I’m on the business side ;) During Basel II a data model was created and maintained during time, but only from a regulatory / risk perspective. My goal is to integrate the sales side into it, which, from a logical point, shall be no real issue, because sale is just the other side of risk, and at least all master data objects are already modeled.

    Best regards
    Rayk

  4. In a previous life I was responsible for defining the Customer Data Model for a large company. When we peeled that back we found we also needed to define models for sales order negotiation (it involved customers), document management (they were sent to and were related to customers, employees & business units (they contacted customers), communications (so we could know who had spoken to whom about what and when) and a few other “things which the business needs to manage”.

    So, we wound up defining an Enterprise wide Conceptual Model for EVERYTHING the business did, needed to know, or needed to know it had done.

    My experience echoes Rayk’s in that I was on the business side. I pushed to get this part of the project started early in the programme. IT wanted to leave it until the final phase, with the Data Migration from legacy sources to this new target needing to be done in that same phase.

    When the Business-side team I lead started on my start-date, we had limited IT support. When we finished the modelling we baselined the model (which the team codenamed “CEIM”, an acronym for Conceptual Enterprise Information Model, and the Irish for ‘step’)it was frozen while we began data migration planning (still about 5 months ahead of IT’s original plan). When we returned to CEIM a few weeks later to review 26 new requirements that had arrived we found that 25 of them were met by the model (base data changes only) and 1 required technology not yet invented.

    Unfortunately by the time we came to start moving data into CEIM, IT has spent our database physicalization and data migration budgets on front-end system changes due to scope creep.

    However, all the business stakeholders were (at least for a while) talking a clear and common language, parts of which eventually made it into data tables and processes but not in the big picture we’d hoped for.

  5. kenoconnordataconsultant says:

    Thank you all for sharing your experience,

    Henrik, I like the sound of ‘Transmodel” the public transportation industry conceptual data model. Is this a proprietary project, or public domain? Is there a link to more details?

    Dylan, thanks for the link to TMForum.org and telling us about shared data models for the Telco industry.

    Rayk, glad you like the metaphors. Interesting to hear about the resistance from IT to developing an Enterprise Wide data model. Did you start “from scratch”, or did you start with an industry template data model?

    Daragh, your experience sounds similar to Rayk’s, in terms of IT reluctance to buy in to the idea of an Enterprise Wide data model. It appears you succeeded without them. Sorry to hear that IT spent all the budget before full benefits could be realised.

    Why do IT functions resist the development of an EDM?
    Why do so many IT functions (and businesses) appear to think “We could not use an industry standard data model, because our organisation/business is different”

    President John F Kennedy visited Ireland in 1963 – in one of his speeches he said “We need men who can dream of things that never were, and ask why not”

    So… If your organisation cannot use the “Industry Standard Data Model” for your Industry, you need to ask “Why Not”. In time, all organisations within your industry will use the standard model – your organisation is at risk of being left behind.

  6. Henrik Liliendahl Sørensen says:

    Link to something about the (European) public transportation data model:

    http://en.wikipedia.org/wiki/Transmodel

  7. Rayk says:

    So, I’ll add another 5 cent.

    Ken, we were not using a industry template. By then the Basel II related offerings by SAP were not of sufficient use to us, so we started with the development of our own model.

    For the IT resistance I may blame myself. Also by then I was part of the BI team on the IT side and our goal was to set up a enterprise data warehouse based on the Basel II data model. What we haven’t had in mind (because of lack of experience) was, that with SAP BW we had a tool at hand not really useful for such an approach. We failed. Greatly. By now any phrase combining “Enterprise” and “Data Model” rings alarm bells at IT.

    For now I see it as part of my Data Governance job to re-establish the working with such a enterprise wide model, and to use it to design IT systems.

    Wish me luck!

  8. kenoconnordataconsultant says:

    (comments on LinkedIn)
    From Scott Ambler

    There are many issues why enterprise models, including enterprise data models, aren’t as prevalent as we’d might like:
    1. Enterprise modeling efforts often take too long, go into too much detail, and become too costly.
    2. The enterprise modeling team doesn’t execute effectively. They create great models, but don’t support them with the teams which should be using them.
    3. The enterprise modeling teams are often ignored by development teams, often because of point #2, but sometimes because the development teams don’t know to work with the enterprise teams. Some stats at http://www.ambysoft.com/surveys/dataManagementAugust2006.html
    4. Lack of senior management support/recognition doesn’t seem to be the primary issue why enterprise modeling efforts get into trouble, but it is a problem. See http://www.ambysoft.com/surveys/stateOfITUnion201001.html for some stats.

    From Krishna Guda
    I tend to agree with Scott, enterprise data models is an expensive affair as much costlier than Enterprise Architecture efforts. CIO may not be held accountable for these things all the time as invariably teams create applications and adopt tools which hit the master data of the organization. This is a difficult sell and probably comes under the spectrum of “enterprise modernization” which generally has a solid ROI already in place.

Leave a Reply

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Follow

Get every new post delivered to your Inbox.

Join 2,374 other followers

%d bloggers like this: