Common Enterprise wide Data Governance Issues – #13. Islands of data.

Islands of data across the Enterprise.

This post is one of a series dealing with common Enterprise Wide Data Governance Issues.    Assess the status of this issue in your Enterprise by clicking here:  Data Governance Issue Assessment Process

In many large organisations, tactical localised projects over many years has resulted in multiple databases of similar but inconsistent and unlinked data (e.g. multiple customer databases).

Historically, business area owners have been reluctant to share “their data” with other business areas – for fear of data corruption, or occassionally for reasons of “empire building”.


– Information Management nightmare

– Multiple versions of “the truth’.

– No ‘single view of customer’.

– Difficulty in meeting regulatory requirements in the area of ‘Know Your Customer’.

– …


CIO to define and implement the following Policy:

  1. New data stores must only be created in consultation with the Enterprise Data Architect, taking into account the target Enterprise Data Model.
  2. Measures are required to replace local islands of operational data with Enterprise wide operational databases, and to implement an Enterprise wide data warehouse.

Does your organisation have “islands of data” – if so, what problems do they cause you?

If not, how did you get rid of them?  Did you build “bridges”?  Do your data islands feed into a central “Information Warehouse”.   Please share your experience.

6 thoughts on “Common Enterprise wide Data Governance Issues – #13. Islands of data.

  1. With apologies to Harry Belafonte:

    This is my island of data
    Where my team has toiled since we tested beta
    I may sail on many an enterprise information sea
    But these data stores will always be home to me

    Oh, island of data
    Willed to me both now and later
    All my days I will sing in praise
    Of your columns, rows, your shining database

    I know there is no way I could ever know
    Another version of the truth that you show
    Never let me believe a data governance nightmare
    But forever dwell in the data I have right here

    This is my island of data
    Where my team has toiled since we tested beta
    I may sail on many an enterprise information sea
    But these data stores will always be home to me

  2. Ken, thanks for sharing, and yes, the evolution of our data architecture should have an approved data model that works within the enterprise an enterprise data warehouse…

    BUT… as enterprises continue to change, shrink, expand and evolve, we will still require those silo-ed data sources.

    ERPs come to mind… I dont think major ERP players will bend their DModels to fit an enterprise, nor will I expect CRM systems to do the same, especially if we introduce internal clouds or external services/clouds.

    Now you hit on a very good theme of having a centralized DWarehouse. I agree that having an aggregated / centralized warehouse of data… with the emergence of streaming your warehouse data for inline analytics, and advancements in Predictive Intelligence, it will only benefit the business to centralize… more wisdom to collect.

    Having governance adopt the new systems like ERP or CRM, will help the enterprise, if only to say, here is a new data source, and its model, and how to access it (or how there will be efforts to mine that data to make it available to the enterprise.)

    Long winded, but I face this every day at all my clients. No wonder the Enterprise Data Architects and DBAs are going crazy… they are getting pressure from Standards of the company (compliance and such), and standards from their vendors(CRM / ERP). Whew… gives me a new appreciation of who they are and what they do…

  3. Jim, I want to see some ocdqvlog’s – vocals and guitar.

    Ken, good policy.

    I think one of the first questions we need to tackle is: Why did the island appear in the first place?

    Lack of support?
    Lack of understanding of data requirements?
    Lack of trust?

    Without tackling the ‘Why’ I don’t think we’ll ever be able to reduce this problem. Do business users really want to be spending time being data producers rather than doing their day jobs? If ‘yes’, perhaps they should consider a new career in IT. I’d say it’s usually because of a feeling a necessity – we need to get to the root cause.

  4. Thanks for the feedback guys,

    Jim – what can I say?

    Garnie – you raise some excellent points.

    Islands of data, and disconnected systems are among the business drivers for implementing ERP and CRM systems.

    Regarding ERPs – When an enterprise sets out to select an ERP, ideally the selection process would include measuring the data models of the shortlisted ERPs against the Enterprise’s “Target Data Model”.

    Regarding CRM systems – CRM systems are intended to provide a “single customer view”. The challenge has always been locating the data required to populate the CRM system. Gartner tells us that 70% of CRM projects fail – uncharted islands of data are one of the primary causes.

    Governance is required – operating above the level of Enterprise Data Architects and DBAs – to maintain sanity, and ensure that no new “disconnected” islands of data are created.

    You are right Garnie, different business areas will always require their own “islands of data” – they may need to contain replicas of master data held elsewhere – that’s fine – so long as the master-replica relationship is understood, maintained and subject to change management.

    Phil – great question “Why did the island appear in the first place?”. I believe the answer lies somewhere in the fact that “Old hardware goes into museums, old software goes into production every night, and old data just slowy degrades”.

    In large financial institutions, customer data may have been gathered 30+ years ago, when a customer opened their first account. So there are many historical reasons why disconnected data islands were created in the past. There is no excuse for creating new ones today – so your questions are very valid in the context of new systems.


  5. Pingback: Process for assessing status of common Enterprise-Wide Data Governance Issues « Ken O'Connor Data Consultant

  6. Excellent post/comments.

    Agree about ERPs/COTS software. But rarely see or hear of this happening. Often the initial cost of implementing the technology is taken as the key factor. Medium/long term maintenance, data management issues etc are often ignored/not even thought about.

    Why do new islands continue to appear?

    1) Very few organisations have enterprise definitions/data model and governance structures.

    2) The incentive model for projects – ‘get in on time and on budget’ – lends itself to a tatical/local/short term approach.

    From a project teams perspective building another island is often seen as a better option:
    – it’s under their control
    – they don’t need to deal with other business areas

    Ways to limit the number of further islands?

    As per the original post:

    Have effective governance structures (is difficult and teams can still circumvent – especially if they control their own budget).

    Provide enterprise wide alternatives – eg enterprise wide data warehouse, data services, integration framework (make it easier to construct the ‘bridges’ when required).

    But these take time to build and you get into the whole issue of who is going to pay, which project will take ‘the hit’ to implement …..

Leave a Reply

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

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

Google photo

You are commenting using your Google 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 )

Connecting to %s