Karen Heath, an analytics solution architect for Accenture’s Health business, posted this question (“Is it really possible to achieve a single version of the truth?”) in this week’s Information Management top stories. Of course I had to run quickly over to the site to read the article – how could I pass up such a red rag!
I have to commend Karen on a fine article and premise. But I wanted to call out a few of her points and explore how really effective they are. I don’t think her article explores these points enough. There are a few such points that always come up in dialog with end users, and some vendors and analysts too!
- Is there such a thing as single-version-of-the truth?
- What is single-version-of-the-truth, anyway?
- Where does this truth persist, and why?
Firstly, is there such a thing? Well there is – and many end user organizations use this phrase, but the reality is that such a thing does not really exist! In fact it’s a somewhat silly question. What most end users really mean is the following:
- Can we at least agree to disagree (what the truth is)?
- Can we at least service our own individual needs with a “view” into what the truth is?
- Can you not push your ideas of the truth onto my ideas?
- Why can’t we just all get along?
The precise phrase we could use is more like “a commonly agreed and accepted set of truths that operate as a foundation, on which we will each derive our own, interpreted contextually centric views”. That is not a very marketable phrase, so we all tend to use “single version of the truth”. Try having this conversation in a Research meeting and you can spend most of the hour arguing about this whole area!
The second question really depends on the focus. In BI land the phrase tends to relate to far more than master data. With respect to the data warehouse, the truth is focused on all the transactions and now content and big data that is collected and used to drive analytics. Master data might make up no more than 5% of the data in that warehouse, at most.
In Master Data Management, the focus is ONLY master data. Or at least, it is for a while. It will always tend to grow over time. See my research round of for 2012!
Now as to the ‘where’ this truth persist – that is much easier to address. Karen’s role seems related to business analytics. This implies something important. It implies (I admit not necessarily so) a focus on a data warehouse of some kind that collects data from many sources, and provides a “single view” (physical or virtual, or even logical) of a lot of data to power analytics and reporting. This is critical to the discussion since, and I have said this before, I don’t think that many organizations are as successful with “information governance” by the (end user) people, for the (end user) people, when the focus is the data warehouse. See “Information Governance on (or in?) the Data Warehouse – Does it Exist?” Mark Beyer, a colleage of mine, sometimes goes further sand suggests that information governance does not exist in the data warehouse for the same reason I site here.
As an ‘ex’ end-user myself, my priority was on the business data within the realm of the business applications I used to allow me to do my day-to-day job. The time I spent on helping IT clean up data in some warehouse to help with reports we get from time to time was a lot less. My bosses, over the years, followed the same practice. The only time the emphasis on the data warehouse made sense was when I had not other choice, or when my business performance required me to do so. Most business performance metrics focus on just that – not data that is not related to performance.
So I believe that trying to instantiate information governance, led by the business, on data in a warehouse to drive reporting, even compliance, is never as east as trying to get information governance established on information in core business systems. If you can get to the latter phase, you will have the attention of the business user – front and center. Thus change management, the level you need, can be identified and measured in a way to motivate the end user to support it.
Therefore the question for me in the article is this – are we focused on governing data in the warehouse (realm of IT), or information in the core business systems (realm of the business user)? They may seem related – we even came up with the phrase, “analytical MDM” to denote the relationship, but the discussion Karen so nicely explores is answered in very different ways depending on the focus of the information in question.
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.