When two applications FROM THE SAME VENDOR can't even maintain a co-ordinated view of the same data between themselves.

Differences due to timing we could live with, as long as it netted out over 24 hours and did represent essentially the same business truths.

Adding in applications from other vendors without factoring this into your integration and process design and you will have no chance

Let’s face it: there is no one version of truth in any corporate data. Any company, even medium sized, has no choice but to support hundreds of applications in production, and tens of them contain this or that version of truth. This is a reality. Ask any manager responsible for operations. Huge variety of apps, platforms, etc. can keep her awake at night.

So, one of the main tasks of a CIO is to create or at least to strive to a consistent and manageable application mix in production, to have sane enterprise architecture. Well, good luck with that!

Moreover, the situation is getting worse every day. There are many business and technical reasons to stick with existing systems in production. On top of that, the new ones are frequently added, often by non-IT departments that push through system after system without any considerations for the cost of support and integration or even understanding of this matter. Mess is the right word.

There may be a hope if a CIO can explain the financial consequences of such short-sightedness to a CFO. But even then the chance is rather slim…

I never liked the phrase "single version of the truth." The closest we come to achieving this is External Financial Reporting. We hire auditors to signoff on the single version of the truth. We also know from the number of re-filing that happen that the truth wasn't readily available.

Information used in management decision making is inherently fluid and doesn't lend itself to the single version concept. The addition on new or additional applications complicates the situation even more.

I think you hit on what I believe is the single biggest challenge that keeps the single version of the truth from being made reality. Decisions are often made using data, imperfect as it is, but what is missing is the structure around how the data was used to make the decision. What assumption are being made, what flaws are likely in the models, documentation of the key ingredients so that the recipe for the latest decision is documented for back testing.

Great article - thanks

While a single enterprise-wide homogeneous, clean set of data is intrinsically a good thing, it's also a hopelessly naive ambition to head directly towards achieving it, at least in the "Big BI" form espoused by enterprise BI and DW tools vendors.

Working from the top down, with the overarching goal of providing a single version of the truth misses the single most important point: all BI is local.

Individual business decision makers are concerned with running their business first and foremost. Their primary need, which needs to be satisfied before any higher level organizational concerns can be addressed, is for high quality business information (intelligence) that lets them evaluate the state of their local business. This local BI is grounded in their local business data.

Of course their business decisions need to be made in accordance with their organization's strategic goals and plans, but reconciling their local needs with the strategic imperatives is their professional responsibility.

From this perspective, the single version of the truth that's the holy grail of Big BI is an emergent property of the consolidation of the high quality local BI efforts. As the local BI improves the improvements can be propagated upwards, resulting in continual improvements in the enterprise's comprehensive BI.

While a single enterprise-wide homogeneous, clean set of data is intrinsically a good thing, it's also a hopelessly naive ambition to head directly towards achieving it, at least in the "Big BI" form espoused by enterprise BI and DW tools vendors.

Working from the top down, with the overarching goal of providing a single version of the truth misses the single most important point: all BI is local.

Individual business decision makers are concerned with running their business first and foremost. Their primary need, which needs to be satisfied before any higher level organizational concerns can be addressed, is for high quality business information (intelligence) that lets them evaluate the state of their local business. This local BI is grounded in their local business data.

Of course their business decisions need to be made in accordance with their organization's strategic goals and plans, but reconciling their local needs with the strategic imperatives is their professional responsibility.

From this perspective, the single version of the truth that's the holy grail of Big BI is an emergent property of the consolidation of the high quality local BI efforts. As the local BI improves the improvements can be propagated upwards, resulting in continual improvements in the enterprise's comprehensive BI.