3 Keys to Better Data Analytics Governance


3 Simple Keys



Over the years, Leopard Business Solutions have  implemented many data warehouses for customers, which in turn  feed data into projects performing very advanced analytics. We have also been on the receiving end of data from customer  data warehouses and data lakes that we’ve used to perform advanced analytics on behalf of our customers.

Our observations are that often analytics is completed in isolation to data governance of data warehouses, which can cause mistrust in the output of the analytics and it can also mean that data can be used in the wrong context if the right people are not engaged properly.

In this paper, we will highlight how disparate governance could be avoided.

What Normally Happens?

As a company, we’ve implemented 10s of data warehouses into various sizes of companies and have subsequently supported these data warehouses with a team of support analysts and developers. Inevitably, a few years after the data warehouse is implemented a request comes through to supply a specialist analyst/team of analysts/consultants with a data dump or access to the data warehouse to be able to pull the data required for analytics. The concept is sound and to be encouraged: specialist analysts providing confirmation, alternative insights or simply a fresh eye.

More often than not, we are the subject matter experts on how the data in the data warehouse was produced/calculated/transformed to achieve the insights/results that are required. However, sometimes we are not included on the project team that’s producing the analytics, the team reviewing the results or helping with the definition of logic that’s being used in the analytics. To a large extent, we provide the data or access to the data and then we have no idea what it’s used for. In some cases, there is a degree of governance in place with other SMEs, but often there isn’t. And this seems to be a common practice in many companies whether it’s with a supplier that supports a data warehouse or with a BI team that is left out of the loop.

This isn’t always the case, sometimes other analysts are brought in and budgets are set before the project manager realises that they need to engage and consult with other people in an organisation. Sometimes it’s assumed that the team that support a data warehouse or the BI team won’t need to be involved and they’re left out of the loop on purpose, but sometimes it’s by omission.

3 Keys To Data Analytics Governance

Here are the keys that we’ve identified as being crucial to ensuring that you have Data Analytics Governance in your organisation:

  1. Collaborate. Collaborate with SMEs and key people that know the calculations/transformations that data goes through before you use it, so they can steer on how the data can be used. Sometimes there are constraints that can create a challenge for this, such as budget. For example, we’ve been involved in some projects where we’ve supported a data warehouse, but there isn’t budget for us to collaborate and spend time with people using the data. This creates a problem when it comes to collaborating and ensuring there is governance over the analytics. This is an example where if it is planned that governance needs to cover Data Analytics as well, then budget can be planned and allocated to cover this sort of collaboration. Often decisions are made in how data is processed, filters that are applied to data when it’s loaded to a data warehouse, what transformation it goes through, etc.. which are all important for the team performing the analytics so that they have the full picture. This is only possible with good collaboration and good collaboration is only possible if it’s intentionally planned and budget set aside for it.

  2. Data Ownership. Having clear ownership of Data Quality and how the data can be used is very important. This needs to be built into the Data Owner’s performance contracts so that the importance of Data Ownership is clearly articulated and understood. The Data Owners will be responsible for the quality of the data and how it is used, which means that the Data Owners will need to be involved in the approval process of how data is consumed, who has access to it, etc. This will  also incentivise collaboration in that the Data Owners will need to be involved in the process of governing how the data is used in analytics.

  3. Tools. When a Data Warehouse is implemented, it’s important to build it in a way that makes it easy for data to be accessed. Rather than people taking a “data dump” from the data warehouse, build the ability for them to connect to the warehouse and use the data as it has been prepared and transformed (e.g. using an API). This allows transparency and means that the data used in the analytics is governed by the same processes that govern the rest of the data warehouse. You also need to build in tools to measure and track Data Quality so that people know the state of the data, performance can be tracked and that the Data Owners know the state of their data and how it’s being used.