During a long period of time up and until the outbreak of the Great War, a gold standard was in use. This gold standard was used to relate the value of one national currency to another: governments agreed that a fixed amount of gold in one country using its currency would purchase the same amount of gold in another country using another currency. It was not a perfect system: it was very much beholden to active collaboration between nations and always subject to external shocks that shifted one nation’s monetary value. But overall it worked, to a degree, when collaboration was active and each country assumed mutually reinforcing monetary objectives. The Great War of course changed all that.
Now think about your business. Don’t you have fiefdoms in place? Each department ‘owning’ their view of the business, using their own data? Don’t parts of the business operate as if they were lords unto their own information use, as if it were their own local currency? And what happens when two departments collide and try to work together? Don’t they argue over whose data is right, or best suited to the task at hand?
As I continue to read more and more about the history of our monetary systems I am amazed at several things. One is that it was less than 100 years ago our leaders and visionaries, including politicians and economists, were only just really understanding the causes of inflation. Gustav Cassel had only just developed a rounded understanding for what was known as purchasing power parity, which is required for a basic understanding of changes in exchange rates.
As I was reading again (this weekend) about the mechanics of the gold standard it came to me that metadata is in fact the key to establishing and maintaining an effective gold standard, or single view of data, across part or all of the enterprise. Actively governing the gold standard meant that nations could trust and use a reliable exchange rate, based on a common conversion rate to gold, between currencies. With trusted and mutually consistent metadata, shared between silos to help reconcile local semantic differences, a data ‘gold standard’ can be realized. If the governance of the shared metadata breaks down, common currency or data exchanges breaks down. As goes the gold standard, so goes the exchange rate and trusted data with it.
So don’t more firms operate with a gold standard? There are a couple of reasons but the most important is that in order to invest in the people, process and technology to support a gold standard there has to be a benefit to someone. No clear benefit, no investment.
A lower level reason but equally a show stopper is that so few organizations really know what it takes to sustain such a data gold standard. For years firms have tried with enterprise data models. Then there was the enterprise data warehouse. Now there are data lakes. All are sold and described as silver bullets when it comes to “all you need for your data”. All of these were sold as “enterprise” solutions, too. A little education will go a long way.
A couple of years ago a colleague of mine (now retired) Michael Blechar formalized the mechanism for creating this data oriented gold standard. He recognized that each silo had its own ‘currency’ in that every solo has, somewhere, is own local currency: or local semantic model. He figured out that in order to share and maintain a consistent semantic model between silos, an active information governance board on behalf of business users would be required. As with gold, this was not a technical solution. And this then leads to the need for a business outcome or imperative in order to obtain buy-in from business users. To emphasize and show how this shared metadata across silos is different from in-silo metadata management, Michael coined the phrase, Enterprise Metadata Management, or EMM. It was a simple but powerful idea. And for the most part EMM has been hard to achieve and elusive. It remains misunderstood by most firms, and also a rare thing. If fact it might be try to say that more mature EIM programs will not likely be sustainable without reasonably active and governed EMM.
A new “topic” is emerging – one called analytic governance. Analytic governance is to analytics what information governance is to information. The outcome is the same (the information, or analytic, is trusted) but the focus is different. And of course analytic consumes data from somewhere too – so there is a dependency. Yet how will he shared metadata from an operational MDM program with an analytical governance program? EMM is the key – not reinvention. And my last idea: Infonomics provides the means to determine purchasing power parity between information silos.