Throughout my career, I’ve worn many hats when it comes to governance, and have had my fair share of learnings along the way – some more memorable (and painful) than others. What’s always struck me about it is that trying to nail data governance can be like trying to catch a cloud.
So why bother? Well, I think there’s a bunch of upside in this space for all of us to realise. More and more of a company’s enterprise value is going to be derived from its data and how it uses its data so it pays to think about what is being wrapped around this precious asset. Good data governance should be a calibrator that helps find the data sweet spot, unlocking how best to use it. It will help you strike the right balance between risk and reward in terms of using and protecting your data. It can help unlock the potential of your data. Getting the right data into the right hands at the right times to inform decisions.
On the flip side, neglecting data governance or letting it just happen can be a real handbrake on realising that dream. What’s the opportunity cost of not using data to its full potential? Is data governance wrongly curtailing that use? What are the hidden sunk costs you are incurring due to remediation on poor data quality or inefficiencies in trying to access the right data? Are your competitors leveraging data better and gaining market share? Data governance is more than data catalogues and perfect data quality. You need to look at the collective efforts of people, processes, and technology, and what will work best in your world – based on your strategy, industry dynamics, ways of working, and culture.
I love technology and there’s plenty of great data catalogues and master data management solutions out there that can help enable data governance. But, at least for this blog, I’ve resisted the urge to focus on the tech. That’s because I believe that it’s the mindset and way of working that leads the way to data governance delight. Getting there is based on being:
I really enjoyed Laura Madsen’s book on Disrupting Data Governance. I found it breaks down the ‘why’ of data governance really nicely and it struck a chord with me when I reflected on when I’d seen some great things coming from data governance activities in the field. It’s a flip perhaps on the traditional view of data governance and it’s a great challenge on how to drive value from your data governance dollar.
Building on that, I’d recommend you tie your data governance efforts to visible, high-value projects where data is integral to solving an important business problem. That will help ensure your data governance is outcomes-focused and will help shine the light on the ROI you’re getting from it. Think it through for every data project you undertake at the get-go and revisit along the way – how’s data governance going to add the most value? What’s the best recipe – the optimal blend of ingredients, added at the right times? In some cases, the data governance value recipe might be two parts quality, two parts visibility. In another, maybe it’s three parts enabling use, one part risk reduction. The project context will drive what it looks like and the value recipe may well change as the project evolves. Here are a couple of examples of outcomes-focused data governance adding value to data projects.
Business Problem: Customer Retention Analysis & Prevention
Retention through data-driven insights leveraging predictive indicators, rather than lagging reports, enables appropriate preventative action
Business Problem: Stealth New Feature Testing
If the primary goal is optimising the use of data to help solve business problems, then it’s the people who will lead this. Data governance in action involves a lot of conversations, interactions, and teamwork. It also requires some real leadership to give it the direction, the air time, and the importance it deserves.One of my clearest memories of this was seeing an awesome executive sponsor absolutely cut through the noise and the ‘no’s’ in a steering committee meeting to get real momentum behind using data in decision making. He constantly challenged his team to use the data, accept it as imperfect but useful, and help improve its quality for the future.
If there’s a shared appreciation of the role that every person plays, mutual respect, and an understanding of who makes the calls in what circumstances, that makes a huge difference. The responsibility for data can’t sit exclusively within the data governance function (if you have one). It’s like trying to make the finance team in your company solely responsible for financial outcomes when everyone can spend. For data governance to really be successful, it needs to be a shared responsibility and embedded in the fabric of the organisation. Find data champions in the business, build strong relationships with them, and support them to continuously learn and share for sustainable data governance. In the context of projects, we’re big fans of multi-disciplinary, self-sustaining teams where data governance has a voice, is working collaboratively with the rest of the team, and is outcomes-focused.
Agile and Adaptive
Data governance has to stay in rhythm with the business and its projects. If your company is committed to working in a more agile way and adapting to change – and let’s face it there’s a lot of change happening – then data gov ops has to emphasise agility and adaptability too. Adapt your approach based on what will deliver the most value in each case. Stay agile so you can pivot, iterate, and change direction at the right time when needed. Also, try to avoid procedures overload when it comes to data governance. I’ve seen it many times where people lose sight of the outcome by becoming too focused on process and compliance. Agility is easier when you’re not weighed down by unnecessary baggage. Think through these 5 things for your data initiatives, adapting, and staying agile in how best to reach the outcome:
What is the nature of the use case? What is the business problem that is trying to be solved and what is its relative priority – what’s riding on this? Clarity around objectives, priorities, and realistic advice on the feasibility of the stated aim of the project at the outset is really important.
Keep the users front of mind. Chances are you’ll have a number of different user groups with different data needs. Adopt a design thinking led approach – engage with users and build out the user personas and their data journeys. In each case, ask yourself how data literate are the users? How can they appreciate the possibilities and the limitations of the data and how best to use it to inform the business decision? How will each of these groups get value from data governance? Will the users embrace the opportunity of using directionally accurate data? Will they see it as a way of refining its quality and supporting its future use? By working closely with the users, you can validate the assumptions early and minimise the risk and uncertainty.
Quality is in the eye of the beholder and data needs to be fit for purpose – but prioritise progress over perfection. Unless you’re dealing with non-negotiables such as laws and regulations, challenge yourself and your users to think about the right balance between quality and timeliness. What is the data quality needed for data to be a net benefit in the decision-making process? If the user is informed and they feel comfortable to use the data as directionally accurate, even if they don’t fully trust it 100%, then perhaps that’s good enough. We do this all the time in our day-to-day lives when we use estimates, approximations, or rounding, so it’s not a new concept. For an end-user, quality can be framed as whether they trust the data to help and know, at least broadly, its limitations. The more use, the more questions asked about the data, the better the quality of data for the future too.
Rachel Botsman, the trusted guru, defines trust as “a confident relationship to the unknown”. If you’re trying to get your users to make that jump into the unknown by using data more, help them out by ensuring there’s a visible trail of the data’s journey back to the source. Let the users trace the data from the source through transformations over time. Data transformations can occur at many places – such as APIs, within ELT tools like Fivetran, and in BI tools like Power BI or Qlik Sense. In other words, bake-in a user-friendly data lineage tool – there are many well-established enterprise tools in the market, as well as powerful open-source options like dbt. Depending on your circumstances, a data catalogue may also be relevant for search over multiple data sources and to help with traceability for audit and compliance requirements too. Look out for our next blog in this data governance series where we dive deeper into some of the technologies out there.
Maybe this is where most people start when they think about data governance and locking down sensitive data. There’s a whole raft of information out there on security and how to get it done. It’s not the focus of this blog, but it is important to always keep it in mind. It’s about using appropriate data, in appropriate ways. That means good data classification practices and a strong working relationship with your infosec team who have to make the calls on access. Enable the right data to be used, for the right reasons, by the right people, at the right times, and consider controls that can detect when this isn’t the case. Adopt a risk-based data classification hierarchy that makes it easier for users to know what data can be used, when and for what purpose so you can strike the right balance between protection and compliance and usability.
Stories from the frontline
At Aginic we get to work with many great clients, delivering awesome data-driven outcomes. We’ve experienced both the devilish and delightful when it comes to data governance in digital transformation.
Here are a few thoughts from our crew on the front line of data projects.
“It’s important to have the data quality conversation early on and use that as a guiding light throughout. There’s a value trade-off between quickly realising project outcomes and getting the data in the hands of the users earlier versus spending more time on refining the data quality to get it 100% accurate. It comes down to what the use cases are and how much appetite the users have for imperfections. Using the data more will almost always help improve its quality over time anyway if you have the right data culture.”
“Data governance often involves a level of healthy tension – between access and security, speed and accuracy, and many other factors. If there is no tension or debate, then one side might be overpowering the other. An effective data governance framework and culture need to enable constructive debates then use the different inputs and perspectives to help decide the right approach for a given situation. It should also challenge trade-offs that are taken for granted – for example, what if we could achieve better speed and accuracy through improved practices and tooling?”
The Wrap Up
If you would like to continue the conversation around data governance, and how you can use it in your organisation as a calibrator to unlocking your ‘sweet spot’, please feel free to reach out at any time for a chat!
Thanks to my awesome colleagues who contributed to this blog. Christine Dixon, Paul Thornton, Viv Shih, Michael Mantfeld, Rowan Walsh, and Angel Chan.
Get in touch with David Hodges
With over 20 years consulting experience including 9 as a Big4 partner, helping clients improve performance and manage risk better. I dived headfirst into the world of data and haven’t looked back! I co-founded Aginic Sydney and love being a part of our awesome growth story, working with our incredibly talented crew to solve business problems differently with data. Love tennis, running, and all things Max Tegmark.Get in touch