Data is increasingly recognised as a critical part of any successful business - with more than a quarter of leading firms now having a Chief Data Officer at the executive level, jumping to over half for firms with more than 100,000 employees. Data governance is a key part of their role, but what that means is not always well understood across an organisation - it can be seen as merely an enforcement tool, telling teams not what to do and dishing out slaps on the wrist when things go wrong. In fact, used to its potential, data governance should be as much about driving success as it is about guarding against failure.
The concept of data governance has evolved significantly over the last decade as firms have got to grips with the challenges and possibilities presented by the ever-increasing amounts of data they’re collecting. Those working outside the field might traditionally think of it as being all about abiding by regulations, setting up rules for what different people within the organisation are allowed to do with different data sets in order that the enterprise as a whole meets necessary legal obligations. The reality is somewhat different - though those things obviously are an important part of data governance, they are only part of it, and the remit of a successful data governance framework should combine that with making sure that the organisation can convert their data into business success.
As Danny Sandwell, technology strategist at enterprise IT solutions provider Quest says, data governance should be about “making sure that you trust your data. Do you trust that you can use it for what you want to use it and it’s the right data and it’s going to give you the results that you want? And do you trust that it’s not putting your company into some sort of jeopardy through that usage? It’s always those two sides - and balancing those things is really what’s changing in the world.” Governance is about understanding the limitations, but also about understanding the possibilities that your data can unlock, especially when combined with the potential of new technology.
Sandwell points to how business interest in new technologies can directly tie into data governance, as we’re seeing leadership desire to “bring AI and machine learning into their organisation and start to leverage those technologies. The ability to do that with a level of comfort and security, is a new metric that people can start to measure. It’s no longer just about letting people know that you have data or looking at data that you have, it’s actually transforming that data from something very prescriptive, into what’s becoming known as data products”.
Data governance has a clear and important role here in helping firms shape these kinds of products, and use this kind of technology, while ensuring they’re complying with regulation and meeting legal obligations. With new technology, as Sandwell says, “the risk is that there’s islands of this starting up in your organisation - who’s in the best place to identify who’s doing what with AI, how can we corral that? Well, someone like a CDO is in a really good position, because a lot of the questions that we want to answer before we feed data into an AI model, are already being answered, for different reasons, by that CDO.”
The ongoing monitoring and deep observability of data quality is essential to ensure not only the initial ready state of the data that will be used in AI efforts, but to also support those teams governing AI models. Both a lack of data governance and a lack of understanding of available data can prevent firms from unlocking the true value of investments in new tech - with 46% of CDOs ranking data quality as the biggest challenge preventing their business making use of generative AI. What might have previously been seen as an issue around analytics and measurement has wider ramifications for a business, but also means improvements to the data pipeline may not have a wider set of impacts that the CDO can point to when making the case for time and resources.
Thinking about data governance as an opportunity, rather than a set of limitations, is key for CDOs and other data leaders to get buy-in from the rest of the business. While preventing potential negative impacts can have a financial impact in terms of what they might cost the business if the worst comes to worst, the challenge is to get that prioritised ahead of other parts of the business that are getting investment because they’re bringing in actual returns. If a CDO can demonstrate that not only is data governance protecting the business, but actively helping driving strategy, then they’ll have an easier time getting the rest of the C-suite on board with what they’re trying to do.
In many ways this is the whole point of of the role of the CDO - where data governance traditionally fell under the remit of the CIO, this was often in the sense of ensuring the business was covered against regulatory issues, and the CDO role grew out of the sense that, as Sandwell underlines: “at the end of the day, you know, data is there for a purpose. And the priority for data is to get the most value out of it. There’s a lot of very forward thinking CIOs that really understand data. But generally, they’re not the right person to be doing that, right, they need that person that has a data-first approach.”
The CDO can and should have a clear strategic role in forming a clear data intelligence strategy. Though firms may increasingly be sitting on top of vast amounts of data, too often there’s little visibility of what that data is, or what kind of impact it could be having on the business. Business leaders should not take for granted that their organisations even have a clear understanding of their data inventory - a 2022 report on data governance indicated as much as 47% of data held by respondents was ‘dark data’, unknown and/or unused by the business. That’s a cost both in the sense of paying to store data you’re getting nothing from, and in the sense of the potential for a missed opportunity.
Beyond understanding where an organisation actually is with the data that it holds, it’s important to have a vision for where it wants to be, and to build a roadmap to get from here to there. It may be the case that too much of your data exists in isolated parts of the business: how can you get from that a proper pipeline where data can exist at the points where it can have the most impact?
For Sandwell a successful strategy has to involve increased data literacy across the organisation: “Just because we say we’re now a data-driven company, it doesn’t mean that there won’t be people across the organisation who are potentially roadblocks to success - not because they don’t want success, but they don’t necessarily understand it.” Part of the roadmap has to be about bringing that change, making incremental steps so that data becomes a part of the culture rather than something siloed in particular areas of the business. A recent survey showed the majority of CDOs are now focussing significant effort on this kind of cultural shift within their businesses, and this is only likely to grow as we see the impact that data and data-driven products can have on a business.
How should CDOs measure that impact? What should their KPIs be? What data should CDOs be tracking about their data? There are a lot of potential metrics around data collection - how important is the data, how much risk is associated with that data, what percentage of data presents compliance issues, and measurements of data quality and observability. But for Sandwell the key is to show business impact: “If you want to get intelligence from your data, you need to have intelligence about your data - and it’s really about turning that into a well-defined, well-curated product, that can demonstrate time to insights and time to impact. So when we come up with a use case for AI, or advanced analytics - how fast can we get there? What does it cost for us to get there? What value did it bring once we got there?”
“Establishing robust guardrails and carefully curating data assets enhances efficiency in use case ideation, data identification, and ‘fit-for-use’ data analysis. Counterintuitively, these measures lead to faster data product delivery and improved data-driven decision-making - investing time upfront with data governance and managing data might seem slow, but in reality it accelerates the data activities at the front end. Governance actually speeds up the entire process.”
If firms are going to take advantage of the evolving data landscape, innovations in AI, machine learning and beyond, they need someone focussed on how these things drive value, and how that value can be demonstrated. In order to succeed CDOs need to demonstrate the impact that properly managed data can have on the business - and that can be as much about educating the business as it is about managing the data: the more visible, the more well understood the impact of the strategy is, the better the case for more resource, which in turn will enable data intelligence to add even more value across the business.
To find out more visit www.quest.com
