Gartner Blog Network


Informatica MDM Day May 12 2014

by Andrew White  |  May 12, 2014  |  3 Comments

I had the good fortune to attend Informatica’s MDM Day.  During my attendance at the MDM day (preceding Informatica World starting tomorrow) I was surrounded with 500 or so fellow MDM’ers spanning many industries.  It was time well spent, hearing end-users talk about their MDM journeys.  Lots of similar issues were heard across the sessions; and though there were stories of success, there were always plenty of issues and challenges remaining.  Great stories from GE and Autotrader.  Enthusiasm is clearly present and much like our own recent Enterprise Information and MDM summit a few weeks ago, information management is exciting and in the ascendancy.  MDM is just an excuse to make information business relevant.   It was a fun and valuable day.

I would also characterize the Informatica MDM Day as the best place and closest point at which Informatica talks about business issues.  I am sure I might see a few snippets of business relevancy in the DI and DQ sessions tomorrow, but the MDM day was much closer to business relevancy.  I like this a lot and wish and hope Informatica does more of such days.  After all MDM, when it is used correctly, is about business process outcomes, not data!

There were some key themes and phrases that I spied in several end-user presentations and stories, and I tweeted many of these.  On MDM specifically:

MDM and PIM

Some of the questions I heard I found thought provoking.  Here was one I heard: “What is the difference between PIM and MDM?”  Of course I know the answer, but a brand new way to respond popped into my head – that PIM comes about due to a poorly designed and incomplete roll out or design for MDM!  The problem was not PIM; it was how the end-user represented what PIM was, using what he said was an Informatica graphic from the vendors’ web site.  It was positioned as a hub in which to store ALL your product information.

MDM hubs are, of course,  meant to only store commonly shared master data, not all data for each master data object.  So why was PIM defined as an uber-hub for all things product master data?  Because it is an expedient design, given the landscape of most organizations.   The data is normally fragmented around the organization, and there is not likely an effective MDM program underway.  Thus this is the first MDM-like program.  Also, ERP is not flexible enough or accommodate all the data needs; and Informatica had not conceived of its DIH until recently.  But with MDM in mind, you would not use PIM the same way now.  MDM may or may not be the authoring application but at least PIM is, for me, a local data integration hub.

PIM acts as a local-application-suite data store that should integrate local information to the core MDM backbone.  PIM is in fact one of the major MDM infrastructure hubs – it is an overweight hub for sure.  If only business applications played well with each other; and if only we could all deploy data integration hubs for our pooled information needs….  Today most local data stores are actually the business application.  This is a hint to where MDM needs to adapt…  Each suite of business applications needs a local data store to serve local needs and subscribe to a central MDM hub.  Maybe Informatica should position its DIH here…with some industry orientation like PIM.  I even thought I heard an end-user imply as much in one of their responses to a question!

Beyond MDM and now information stewardship- more than a dashboard

I heard several times on stage: “MDM was traditionally a headless application”.  Actually, MDM was not.  The speakers were actually referring to CDI that preceded MDM.  CDI was in many, but not all, cases, headless.  CDI was a technology solution to a business problem.  It was missing the business dimension and that is one reason why MDM came about.  MDM, once it was defined, had a full blown UI that CDI vendors had little understanding of, at the time.  That is because CDI alone was not enough to make MDM real.  PIM was needed!  PIM had the UI, had tackled authoring, but had avoided matching and data quality.  Data Director was Siperian’s solution at the time on top of a CDI hub.  Even Initiate developed a UI in response to customer requests to author data in its customer (registry) hub.  Some vendors in those days used the term stewardship, not really knowing how important that name would become.  See below for more on this issue of what stewardship really means.

But let’s not quibble about the past.  Let’s look to the future.  The current weakness with most MDM solutions is exposed if you ask, as we heard through the day (I paraphrase):

  • What is the point of an information steward role?
  • What work do they do?
  • How long does it take?
  • Is it different to what normal end user do every day?

You only have to ask three software vendors to collect 5 different response to each question!

For Informatica the same user interface (Data Director) is offered whatever the role – steward or end-user.  Yet the role of steward requires functionality found in other Informatica products (so far) with some present in Data Dirctor.  This perspective does not align perfectly with Gartner’s.  Our view is clear:

  • Most business end users do what we would call data maintenance.  This includes maintenance, and day to day work such as adding records, matching records, processing data quality routines
  • Information stewards are business end-users who monitor the state of information in their preview and enforce information policy

And there should be far fewer information stewards than there are regular business application users.  The first, larger pool of users need authoring tools, with data quality tools to do matching, with workflow controls  etc.  They may author customer data or product data, I don’t care.  Stewards do NOT author data; they monitor and measure and generally solve the exceptions that cannot be handled through automation.  Stewards do NOT match and merge customer records – normal users will do that.  Stewards solve the problems that cannot be solved normally.

Thus a new solution is needed for information stewards, with similar capability to MDM solutions but with some differences. See my blog on this just 2 weeks ago.  Informatica is not on the exact same trajectory, yet, but is close.  I have hope the Informatica message on this may clear, soon.

And now for something completely different (well, not really)

At the same time there was for me, an new underlying theme that bubbled away in all sessions.  Every end user story included a discussion with a need to “document and understand the flow of business information across business systems”.  This understanding includes:

  • Publisher of information
  • Subscriber/consumer of information
  • Business rules applicable
  • Business users of the data
  • State – past and current – of data in question
  • Timeliness of information flow across systems

And all this was primarily at the business or conceptual level, not at a n IT or physical level.  This latter level can be generated with traditional IT tools.

Several end users reported using spreadsheets and other tools to help document much of this.  I know several metadata management tools and information architect tools can also help with discover or design/documentation.  But these are all very IT centric!  I think this will be another area for MDM and IG vendors to explore with additional investment.  There is plenty of IT tools in this space, but not many (any?) business focused solutions…  And MANY end-users are figuring this out….slowly…

How to Mix Messages

I did hear one speaker suggest that MDM emerged because firms tried to improve dimensional management in a warehouse powering analytics.  I beg to differ.  It maybe that a vendor from the past that has MDM DNA may have done just this.  In fact there are several.  Kalido was one.  Hyperion acquired one (Razza, though they had customers in ERP as well as BI too) before they themselves were acquired.  Microsoft acquired one.  But the focus on “analytical MDM” is not where or why MDM really formed.  What we refer to as MDM really emerged because the operational data in our transaction systems was not governed and the warehouse was just one place where this was visible, but not the only place.  In fact at Gartner 90% of our MDM research and end-user inquiries are oriented around operational data in transnational systems; 10% or less is focused on non-operational data.  During this session the audience was polled for who is looking upstream in operations (60-70%) and who is looking downstream at analytical data (30-40%).  Good luck that last bunch.  Few in the business have any motivation to work with IT on an information governance program on a data warehouse.

And finally folks…

To wrap up my first Informatica MDM day.  Here are some crazy ideas for next year.  I felt the need for some upfront vision session to kick things off with MDM and IG in the future…

  • Dennis SVP and GM, paints a vision for the firm (any industry, end user scenario) 3 years hence.  What is the dream?  What is possible?
  • Rob Karel,  VP MDM Product Strategy, changes his title to VP, Information Governance Solutions product strategy (including MDM, data quality, semantic modeling) presents a vision of the new product and solution capabilities needed to meet these future requirements.  This is the IG and MDM road map.  Rob also demos a mock up of the pending new release of the Information Stewardship solution that would integrate with Informatica MDM hubs, DIH and any other vendor hubs for that matter.  The new Info Steward application will also demo well with content being governed, as well as structured information.
  • Ben Rund and Ravi Shanker, joint VP’s MDM industry solutions, reporting to Rob, chairs several sessions that include demo and end-user case study for each of the industry specific MDM solutions, built on the new common platform of information capabilities; omni-channel information hub (was known as Informatica PIM); supplier procurement hub for MRO, patient hub, pharma manufacturing hub, etc.

 

Looking forward to Informatica World tomorrow!  See you there.

Category: content-governance  customer-data-integration-cdi  data-governance  enterprise-information-management-eim  informatica  informatica-world-14  information-governance  master-data-management  mdm  mdm-of-customer-data  mdm-of-product-data-mdm  product-information-management-pim  

Andrew White
Research VP
8 years at Gartner
22 years IT industry

Andrew White is a Distinguished Analyst and VP. His roles include Chief of Research and Content Lead for Data and Analytics. His main research focus is data and analytics strategy, platforms, and governance. Read Full Bio


Thoughts on Informatica MDM Day May 12 2014


  1. Ravi Shankar says:

    Well written blog Andrew! It was nice to have you at the MDM Day yesterday.

    Thank you for the compliment about the MDM Day bringing out the business value through several customer stories. During our lunch yesterday, you explained your answer for “How MDM and PIM are different”. It was very insightful to read how you articulate the difference with DIH in the mix. Your point about needing a new UI for Information Stewards is well noted.

    I would like to extend one clarification regarding analytical MDM: our users are not using our MDM to correct the structure for dimensions in the DW; they are simply using our MDM to correct the dimensional “data” within DW, just like they are correcting the master data within transactions in operational systems. That distinction makes us different from Raaza, Kalido, etc.

    Thanks for your advise regarding next year’s MDM Day. We will incorporate those suggestions. Enjoy the rest of Informatica World!



Comments are closed

Comments or opinions expressed on this blog are those of the individual contributors only, and do not necessarily represent the views of Gartner, Inc. or its management. Readers may copy and redistribute blog postings on other blogs, or otherwise for private, non-commercial or journalistic purposes, with attribution to Gartner. This content may not be used for any other purposes in any other formats or media. The content on this blog is provided on an "as-is" basis. Gartner shall not be liable for any damages whatsoever arising out of the content or use of this blog.