Why data governance is like Fight Club

When building a data-driven culture, there are some things you don't talk about, says Vattenfall's governance lead

"The key to data governance is not to talk about data governance," said Sebastian Kaus, engineer and data governance lead at energy company Vattenfall. "It's like Fight Club."

Kaus likes a metaphor (although he admits he did not coin that one). Metaphors, similes and stories reach the parts that the deadening jargon of governance cannot. They come into their own when building a data-driven culture in a company like Vattenfall, whose 20,000 staff are spread across several European countries as well as in its home turf of Sweden. The inevitable language- and role-based silos that exist in an organisation of that size and spread mean a shared vocabulary is essential. But if you want to foster cross-departmental collaboration based on high quality data - a must for 100-year-old Vattenfall with its stated goal of making 'fossil-free living possible within one generation' - whatever you do, don't mention the G-word.

'Governance' smacks of edicts handed down from above and dusty documents that no-one reads, but a thriving data culture must be grown from the bottom up. And that starts with listening to people, said Kaus, and particularly to what they're grumbling about.

Sebastian Kaus: 'Don't talk about data governance'

"People are always complaining about things: this thing isn't working; this thing takes me three days; I can't do this because my colleague is on holiday. We pick up those things and say, okay, well, we want to do it for you, so let's do it together."

Doing it together means that Kaus's team helps them define the issues and the surrounding vocabulary, and then, with their guidance, implements the sort of solutions that can create a positive momentum.

"Involvement creates commitment," says Kaus. "And you need to have a lot of committed people, because culture is when that translates into something. When people say, 'okay this is how things are done', then you have a culture, right? Like putting a tree in your living room over Christmas is part of culture."

With senior management, the approach is a little different. Here, the trick is to turn the data governance problem into a business problem, and then work it into the digital transformation.

"The head of asset management is interested in managing assets and not in digitalising asset management. And they are not measured by in terms of digitalisation, they're measured by EBIT or whatever. And when we can connect the digitalisation problems to the existing business problems we have a very good incentive for the senior management, for the decision makers."

Building a data catalogue

The vernacular vocabularies around these changes are stored in a searchable data catalogue, the idea being that this becomes the first port of call when anyone has a digital or data-related question. But this doesn't just happen. Building a critical mass requires some serious groundwork, appointing data governors and incentivising people to contribute, but if successful that's when bridges start to form between departments, countries and functions almost organically.

"People slowly but steadily get to know each other, and they learn they can type in something in the search bar and figure out 'oh, this is already happening in this other business area'."

As well as individual contributions, all public data on the Vattenfall data platform is also synchronised with the data catalogue, which acts as a front end to all the company's vast repositories.

Technology-wise, Vattenfall has used a data catalogue and associated tools from Alation for a year, having trialled it first. It was chosen over alternatives like Collibra, which were found to be too complex and enterprise-oriented for the company's needs.

"It's way too big and also way too techie. Alation is the sweet spot between having those low-level technical capabilities and usability for adoption, so you don't need to offer a lot of training. It looks like a webpage. If you can read, then you're good to go. Of course, there are always things you'd want different, but Alation's a rising star, and we see it as a partner that we can grow with together on this journey."

Kaus's team is currently focused on improving data quality through automated checking and assessing. After all, no matter how good the platform, garbage in garbage out always applies. To return to his Christmas tree analogy, once the culture of having a tree in your living room has taken hold, the next stage is simplifying it by organising a tree delivery service to the home.

"Quality is key to having trust in your data in the end. We have now established procedure and people are speaking the same language, so now it's about lifting it up to the next level, applying our knowledge and our vocabulary in a consistent automated way."

No matter how you dress it up, data governance, will never be a glamorous topic. But perhaps the second rule of data governance should be 'make sure you explain the nature of the commitment to the board'. With its necessary data cleaning and community building, the reality can be a long way from the slick vendor slideshows aimed at CEOs.

"The nitty-gritty stuff, the bread-and-butter stuff - it's not fancy, but nobody likes to talk about that," Kaus said.

"But it's like a dishwasher: if it breaks the restaurant will not work. So we need to try to convert those not-so-fun topics into something that is meaningful to the people."