It’s time to start retiring the notion of “one version of the truth”.
We’ve made some progress…
Most organizations are taking concrete steps towards eliminating the obvious and pressing problem of erroneous, inconsistent data, with increasingly tightly-integrated data quality, data integration, metadata, master data, and semantic technology
But most (all?) organizations still have a long way to go before employees can confidently use the same terms to communicate the same concepts. And it will be even longer before there’s a common view across business ecosystems of partners, customers, suppliers, and regulatory authorities.
…but ultimately, the goal isn’t achievable
Paradoxically, the closer we get to providing “one version of the truth”, the clearer it will be than no such thing exists.
In today’s fast-moving, complex world, striving for “one version of the truth” is like striving for “one version of politics”, in that:
- It’s crazy to think that we’d ever get there, given the natural divergences of opinion over what is happening, what the priorities are, and what should change; and
- It’s not at all clear that we want to get there (the company as a totalitarian state is not an appealing vision) — diversity of opinion is vital for survival
Facts are still important…
Clearly, this doesn’t call into question the importance of collecting better “facts” — in business as in politics, one of the biggest problems is that many policy debates happen in a vacuum, without looking at what’s really happening on the ground.
A good example of how things are improving is the National Offender Management Service (NOMS) in the UK. There are timeless, perfectly reasonable disagreements about the priorities of different roles of the justice system (dissuasion, retribution, rehabilitation).
NOMS is using business intelligence to provide an integrated view of trustworthy information for each offender. By making more informed, fact-based choices about things like sentencing guidelines for different offender profiles, they hope to cut re-offending rates by 10% by the end of the decade.
…but there can be many interpretations of them
Alas, the best facts in the world are subject to different interpretations. And there may be different definitions of “success”, and the routes to get there.
To pick up the NOMS example: there’s a body of evidence that shows that “outward-bound” type courses result in lower re-offense rates than custody sentences for young offenders, and are typically much less expensive. But the courses are often unpopular with voters and politicians, who see them as rewarding criminal behavior with “fun” opportunities that are not available to non-offenders. And there’s no obvious right answer to who is correct.
The same doubts applies to business. Was the latest initiative a failure because it was a bad idea, because it didn’t receive enough resources, or because the wrong person lead it? It is often impossible to tell definitively from the data alone.
The real challenge for information systems in the future is the ability to cope with this ambiguity, to allow all the different “versions of the truth” to co-exist, while still providing enough direction for the organization to more forward in a coordinated way.
Lots of work to do in the future
We can currently only glimpse at all the future changes that will be required, but here are a few of the subjects that BI/PM vendors will have to tackle in the years to come:
- More explicitly allowance for “fuzzy” data and margins of error
- Support for multiple hypotheses, that can be tested as new data becomes available
- Much better systems for collaboration around BI
- The use of more advanced strategy management systems that go far beyond simple dashboards and scorecards, and allow every level of the organization to participate actively in the process of gathering information, and contributing to strategy.