Start a Discussion
BI
user-pic

Why is a Single Version of the Truth Still Difficult to Achieve With BI?

Vote 0 Votes
From Nari Kannan: In terms of BI, why is a single version of truth still difficult to achieve?

5 Replies

| Add a Reply
  • Every company still has multiple ERP systems, multiple brands and instances of databases, and siloed systems. Even BI tools themselves are scattered. Spreadsheets still reign as the BI tool of choice!

  • A single-version of the truth by definition means that there's a single representation of critical data such as customers, products, assets, and so on (a.k.a master data) that's unique, complete, and consistent, and becomes the most reliable and authoritative information for the entire enterprise. With data duplicated, inconsistent, incomplete, and spread in silos across disparate systems, it requires a special technology to recognize (disparate data), resolve (into a single version of the truth), and relate (them to derive some meaning). In addition you need to persist the single version of the truth, store the complete history of all changes, and lineage of how duplicate data has been merged. BI systems, or the ETL systems that extract, transform and load data into the BI systems, contain the technology to create the single version of the truth as defined above. Master Data Management (MDM) is growing discipline within enterprise software space that provides the unique value for creating and maintaining a single version of the truth that BI, CRM, or ERP systems are ill-equiped to handle. That's why a number for companies in various industries use MDM to feed reliable dimensional and hierarchical data to BI sytems for compliance reporting, effective sales/ marketing/ management reporting, etc. Here are some good articles on MDM and how to use it with BI - http:\\blog.siperian.com.

  • I agree with both Joe's and Ravi's comments. I will add one BI-centric reply . . .

    Even within classical BI systems, more than one version of the truth can persist, unfortunately. Example: multi-dimensional analysis tools (OLAP, in the past) require data to be moved into yet another data store (an OLAP cube, for instance) where the business user immediately ascribes suspicion around its veracity. New server-based, multi-dimensional techniques avoid this problem by integrating analytical views "in-memory" so original source data systems can be used to satisfy query requests rather than moving the data into a cube. These in-memory views persist only as long as they are needed and can be a very efficient way to deliver multi-dimensional analysis where the costs and complexity of old-fashioned OLAP may not make sense.

    Eventually, a combination of new technologies mentioned in these replies will help organizations overcome the multiple version of the truth issues.

    Brian Gentile
    Chief Executive Officer
    Jaspersoft

  • Getting a single version of the truth is difficult because of technical, cultural and process obstacles. Process in that most companies as they grow have multiple processes which sometimes overlap in their need for kernels of data. And in the absence of readily available depots to draw from (i.e. self-service IT), they create it themselves by spreadsheet or any means necessary (the train cannot stop!)

    Then we have the phenomenon of dispersed organizations and dispersed functions and dispersed roles that create silos of data that again must be rendered for personal use. The nature itself of information and its variants is that it is self-replicating like a virus and grows and recombines in forms few can anticipate. Especially outside of traditional legacy data (financials, production, etc.).

    So the answer to this dilemma can never be just technology. It has to be cultural which implies process and behavioral changes that are mandated from the executive suite. A more disciplined approach to the creation, recombination and distribution will get us closer to the one version of the truth.
    This is something we have been working on for a couple years at Exact and we still find company culture (i.e. how things are done) trumps technology platforms even in the best case scenarios.

  • Perhaps we need to revist what we mean by "single version of truth"

    It's important to realize that in some businesses a single version of truth for some master data elements may not be completely achievable. There are sometimes real business reasons to keep multiple views of a master data element. In the pharma industry it may be necessary to create a unique view of a customer from a contracting perspective that may not align with the view of the customer that field reps see.

    As you are defining your master data elements, it's important that you choose a master data solution that has flexibility to support multiple master data views if there is a valid business need.

    In the end it may not be about a "single version of truth" but rather having a set of controlled processes and technologies that allow you to maintain, track, resolve, govern and relate multiple views in the appropriate context for your business.

Add a Reply

Recently Commented On

Monthly Archives

Blogs

ADVERTISEMENT