BI data governance

What to expect for data governance in 2020?

Data governance always has been a complicated issue for most organizations. That won’t change in a big way in 2020. In fact, the increasing prevalence of technologies like artificial intelligence (AI) and machine learning (ML) may show up some of the pains even more. Don’t take that to mean that companies aren’t becoming more mature in their approach to Data Governance, though.

AI, ML, the Internet of Things (IoT), and full process digitization will be a focus for organizations in 2020. Companies see them as required capabilities in the future and so are willing to invest in more digital innovation. 'This is expanding the governance lens and I’m seeing AI Governance becoming a reality in leading organizations', said Kelle O’Neal, founder and CEO of First San Francisco Partners. This trend shows that companies are seeing value in Data Governance so they’re extending successful practices into other areas of their business, she said.

Organizations are realizing that AI is only successful when built upon a solid data foundation, thus driving the need for data governance, agreed Donna Burbank, managing director at Global Data Strategy:

'I’ve had venture capital organizations approach us to train their AI startups in the foundations of data governance as a condition for investment', she said. 'I see that as an extremely positive sign pointing to the widespread need and adoption of data governance principles'.

And yet poor data quality resulting from problems with data governance bedevils AI and ML outcomes and there’s no sign that that won’t be the case next year too.

'Artificial intelligence and machine learning have been way oversold. Data quality gets in the way of getting good results and organizations spend way, way more time cleaning things up', said Thomas C. Redman, Ph.D., 'the Data Doc' and President of Data Quality Solutions. He estimates that more than 80% of AI and ML programs continue to fail because of this.

Governance defined …Yet?

One question that many companies will continue to grapple with in the new year is figuring out just what data governance is. In simple terms, said Redman, it’s a senior oversight function whose leaders advise the board or senior management about whether a data-related program is designed in the best interest of the company and is operating as designed. And as he sees it, no one is doing that yet.

'There’s all talk about data as the most important asset, but having that oversight level would be essential if that statement were correct', he said. It’s not about plugging in various tools but about thinking of just what data governance is … and what it isn’t:

'The term ‘governance’ is being used for everything from moving data from here to there to something about how you operate analytics. That’s not the proper use of the term'.

Getting roles and responsibilities right is critical, he said. Data governance should be business-led and IT supported, Burbank remarked: 

'All areas of the business need to have accountability for the data in their domain and establishing data stewardship roles is critical to ensuring accountability at all levels of the organization from strategic to tactical'.

Chief Data Officer (CDO) roles are becoming more common, and the office of the CDO does best when it reports up through a business function like operations, strategy, or shared services, said O’Neal, or even finance if that team is influential in driving enterprise programs that result in corporate growth.

Organizations that have matured their data governance practices will grow from a program culture to a data culture, which is one:

'Where new employees start learning about data governance as part of their new-hire training, and data governance and management are part of the conversation at the board level', said O’Neal.

What will data governance look like in 2020?

It’s true that there haven’t been drastic changes in how far we’ve come with data governance over the past year, but O’Neal finds that companies are showing progress:

'More and more companies are moving from ‘what is data governance and why should I do it,’ past creating a strategy, into not just implementation but also operationalization, where their data governance is really embedded with other project, decision-making, and ‘business as usual’ operations', she said.

In terms of a formal, structured approach, the DAMA DMBoK is gaining wide acceptance, which is a positive step in aligning best practices, Burbank said:

'While data governance is certainly not a ‘cookie cutter’ approach that can be simply taken from a book, the DMBOK does offer a good foundation on which organizations can build and customize to align with their own unique organizational needs and culture'.

In 2019, Global Data Strategy supported data governance for a diverse array of sectors, including social services, education, manufacturing, insurance, building, and construction. 'It’s no longer just the traditional sectors like finance who understand the value of data', she said.

Big value in small wins

It’s really hard to impose Data Governance frameworks on big data at enterprise scale. It is better to start with small data first and Redman is optimistic that more companies will do so in 2020.

'Practically everyone sees the logic in small data projects', he said. 'Suppose that only half of a hundred small data projects succeed, that’s a huge number of wins', with positive implications for cost savings and improvements in areas like customer service. And solving more of these leads to learning about what it takes to solve big data problems. 'If you build the organizational muscle you need doing small data projects you can tackle big data projects'.

Following the classic rule of thinking big and starting small in order to have the proper data governance framework and foundation in place is what works, Burbank said. Establishing small 'quick wins' shows continual value across the organization.

Tools to help

2018 saw astounding growth in the data catalog market, O’Neal said. Data catalogs provide information about each piece of data, such as location of entities and data lineage. So, if you haven’t thought about that yet, it’s time to do that this year, she said.

The good news is that the modern tools for Metadata Management and data cataloguing are much more user-friendly and approachable, according to Burbank:

'Which is a great advancement for giving business users self-service capability and accountability for metadata and governance'.

Redman noted that 'you can love your data governance tools, and I do too. But if you approach the problem wrong it doesn’t matter what tools you have'.

What’s up next

In 2020, the organizations that are able to get their own data governance in order will reach out to others in the industry to establish cross-organization data governance and data sharing agreements:

'For example, organizations in the social services or medical arena are looking to provide cohesive support for individuals across organizations that provide the best level of service, while at the same time protecting privacy', Burbank said. 'It’s an interesting challenge, and an area of growth and opportunity in the data governance space'.

There’s an opportunity this year for companies that are moderately mature in data governance to think about how to embed practices in business processes and decision-making structure of the organization. Places to look for embedment opportunities, O’Neal commented, are new project initiation and project management, investment approval and funding, customer creation and on-boarding, product development and launch, and vendor management/procurement.

Expect data analytics and BI to continue to be large drivers for data governance:

'As more organizations want to become data-driven', Burbank said, 'they are realizing that the dashboards used to drive business decision-making must be well-governed and well-understood with full data lineage, metadata definitions, and so on'.

Author: Jennifer Zaino

Source: Dataversity