This post follows a recent Red Olive round table, at which technology leaders from the insurance sector discussed data governance. Here, we follow up on their thoughts about implementing effective data governance policies with an organisation.

Data needs to be managed. Doing this well, wherever it is in its lifecycle, however it’s stored, and for whatever reason it’s being used, can only be achieved with effective data governance.

It has become clearer over the last two decades that data is an organisation’s most valuable business asset. It underpins every decision, is the basis on which it forecasts trends and performance, and helps it map decisions to outcomes. Its management might be overseen by a CDO and systems team, but the benefits of that data are evident company wide, so any data governance policies will need company-wide buy-in, if they’re to be effective.

Back to basics

Getting this buy-in may mean starting from scratch, and explaining to the business what data governance is, even before outlining any potential benefits.

After that, a focus on those aspects that will deliver the quickest return is often most effective since the business side of the organisation will usually only appreciate something if it delivers a tangible outcome. In short, get some quick wins before moving onto the bigger picture. Anything else will require more significant resources and buy-in, both of which will be hard to deliver without demonstrable benefit.

Bring the company together

It’s important that everyone speaks the same language. This means implementing a consistent set of data definitions, so that everyone can come to the table with a single version of the truth.

By wrapping business processes around this agreed version, minor differences between departments will no longer matter: they’ll each be able to talk with and understand one another, share a common set of data, and derive the same view from it, even if the underlying terminology differs.

This may require the use of a semantic layer, which decouples the structure of data from the way that it’s discussed. This is just one aspect of implementation on which the organisation will need to lean heavily on its engineers.

Shortcuts, not cutting corners

The processes of effective data governance can often be seen as a barrier to effective, efficient data use. This can be overcome by automating tasks, like cataloguing or appending relevant metadata, wherever possible.

An organisation therefore needs to instil a data governance mindset within its engineers. Doing so will help them to understand the benefit of the automations they’re implementing within the codebase, with the aim of ensuring consistency within the data itself.

Automation reduces the effort of working with incoming data. As soon as it’s gathered, it can be logged and processed in a logical manner, which immediately increases its value. It can be checked for consistency and compared with accepted tolerances, and dashboards can be used to highlight when the data strays beyond set thresholds. That would indicate either that there are some disagreements that need addressing, or that the data should be excluded from business processes, at least in the short term.

Data about data

When we’re thinking about data governance, therefore, we need to think about what the data means, where it flows, where it sits, and what it’s used for. Lineage mapping plays its part here, tracking the data from the point of collection to the moment of consumption and recording every change or interaction along the way. It shows how those interactions have affected an outcome and will feed back into any quality score applied to the data at rest.

Once you understand what a ‘good’ version of your data looks like, you can more effectively write and implement rules to deliver it. And, once you have those rules in place, you can minimise the likelihood of working with ‘bad’ data, or of introducing risk by using the data inappropriately or mismanaging its flow through the business’s processes and systems. This is key to effective data governance.

Data governance ensures that any decisions made on the back of that data are fair on the customers and stakeholders concerned – and, equally, that they’re valid. The second of these may well be the main concern of internal users day-to-day, and a focus on that aspect may be most helpful to IT and the CDO in selling the principle of data governance to the wider business, but ultimately the aim of applying data in a fair and ethical way should be the primary concern of data governance.

Data governance and data management

If a business is to implement effective data governance, it must do so in a logical, target-driven manner. It may, for example, set itself a target of understanding all its data, or remedying a specific number of issues, by a fixed date, and progress against these targets can be monitored and assessed.

Setting targets keeps it focused, and can keep regulators happy, too, since they often focus more on processes than on the data those processes impact.

However, a balance must be struck. It’s important not to focus entirely on regulation, to the point where the needs of the business are overlooked. What’s required is an integrated workflow within which the organisation can consider data governance as a part of its ongoing activities.

This is almost impossible to achieve without involvement across the organisation, so it’s not a campaign that could be driven by the CDO or CTO alone – the ‘business’ must work with them to formulate and implement data governance principles that work for all parties: management, staff, customers and so on.

This can be achieved by demonstrating to stakeholders that the result will be of benefit to the company itself and potentially to them, and it’s not something being introduced purely for compliance with the law.

Indeed, when done well, compliance can be a by-product of effective, efficient data governance that was otherwise implemented to serve the needs of the business.

Whether you’re just getting started on your data governance journey, or you want to refine existing processes, get in touch. We’ve partnered with businesses in diverse sectors – not just insurance – to help them implement the latest processes and technologies to realise the potential of their data.