The Analytics Governance Pyramid
Written by: David Arenbo, CEO
With ever growing data at disposal, and an increase in skilled and motivated employees, the need for some sort of analytics governance is as important as ever.
There are several factors that must be considered to set up guidelines, and they have to be balanced between data, tools, skill of users and cost/benefit.
In general - data that is used for financial reporting is at the top of the governance pyramid (see Figure 1), where data needs to be accurate and of high quality. You probably want to make sure that reports that are built also are “locked” so no one can change them and that the data is certified (lots of tools provide that type of functionality.
Figure 1: The Analytics Data Governance Pyramid
At the other end of the governance spectrum you will find people who have the skillset to collect and integrate huge amounts of internal and external data themselves. These guys will probably also apply some fancy algorithm (maybe downloaded from CRAN) where the purpose is to evaluate a theory, do a one time analysis, or check trends, and where a small group of people with insight to the challenges at hand will use the result of that work.
The challenge is to identify the different layers (there are usually more than two) that your organization needs to fit into that pyramid, what method(s) to use to ensure that the level of quality is kept and last but not least make sure that there are enough resources to keep it up to speed.
Informed and data driven decisions are here, and your organization must get a handle on it before you end up in “my report is more correct than yours because…”