How to change data culture within a government agency, one click at a time

This article is written by Milda Aksamitauskas, Chief Data Officer at Wisconsin’s Department of Justice

• The problem: People are rightly wary of seeing significant changes to their data culture
• Why it matters: Organisations need to adapt to the digital future
• The solution: Going slowly, rather than foisting change on users, is vital

I first experienced the possibilities of the internet when people still connected to the internet through a dial-up connection and you needed a lot of patience until information loaded on the screen (for those feeling nostalgic, here is a video). I didn’t have much money so I switched between internet providers after a one- or three-month free trial would end. The memories of loading information from disks and the “Wait, still loading” box are vivid in my memory.

• Want to write for us? Take a look at Apolitical’s guide for contributors

Working on data management initiatives in government, I sometimes feel that the patience level required to sustain and grow a data culture within an organisation requires the same patience as browsing the dial up internet.

Once managers start talking about data-driven decision making within an organisation, usually an overnight change is expected. The next morning in the office, everyone will be typing in searches for their data needs and accurate and informative answers will start appearing on a screen. Actually, the next morning will remain the same as yesterday if incentives and motivation to use data are not changing.

How to champion change, responsibly

I want to share several practical steps that data change champions in government agencies can take to bring data to the decision-making table. First, ask ‘What data do we have?’ It seems like a simple question, but in organisations beginning the data journey, a list or catalogue of data assets may not exist. There are likely to be people who have worked in your organisation for many years and they know a lot about data, but this knowledge is not written down.

In my experience, enlisting an old-fashioned pen, paper and interview plan with people who work with data at any level works well. There are many technology companies selling tools that help create data catalogues, but the initial round of changes starts with conversations. In my experience, a standardised data catalogue interview works well. First, I developed a standard list of questions that everyone got. Data definition was broad and it really helped. I learned that we had databases with real-time data updates and boxes of filled out paper forms. Excel spreadsheets are also used to produce, collect and analyse data. Interviews provided some expected outcomes: I met a lot of people, learned about fascinating roles and challenges they have, and picked up on some common vocabulary people use to describe data.

I continued the mantra of “I am just documenting what we have, no judgment on how it is done.” The unexpected outcome was that people kept coming up with other datasets during the interviews and we would schedule follow up conversations to take notes about those newly-remembered data assets. Over nine months of work at about 10-20 hours per week,we documented close to 200 data assets and completed the first round of a data asset inventory.

Find out who are the data owners

During the data catalog project, I introduced the term “data owner.” My main goal was to get names of actual people. I encountered quite a few debates about the definition and whether data is owned by programme staff or the highest official in the organisation.

A strong data governance programme is the foundation for further work on building a culture of data-driven decision making.

I stuck with my initial task to put names of actual people who work with that particular data. Seeing your name typed up in a spreadsheet brings importance and responsibility. The data owner question was not popular. But I did observe a shifting culture over time as “data owners” are now a much more common term used in discussions. And it is agreed that data owners know most about their data so if you have a question, you should ask them directly.

Who is making data decisions? In other words, what is the frame of data governance structure in your organisation? This may be a new term for your organisation, but data governance is always happening. Most likely, data decisions are made by a small group composed of staff well familiar with the data and context of questions asked by the requestor. The requestor usually has no appeal rights if they are not happy with the result.

Engaging political appointees in data requests usually is reserved for very rare cases. A more mature data organisation establishes a more centralised data governance structure with more parties involved in the decision making. Transparent data governance is difficult. At the heart of the issue is power. Transparent decisions about data governance mean decisions are moved away from individual people or small groups to a larger group and more open conversation. The power dynamics change. One lesson learned for those implementing a formal data governance is to be prepared to articulate the vision and need for data governance - and keep educating everyone why it is a better way. A 30-minute conversation that moves a needle may require hours of preparation of talking points, checking on your own emotions, a long list of examples and your top-notch appeal and persuasion ability. Practice with a trusted colleague is very helpful.

Good governance

Many organisations create data governance groups. The names used may include board, committee, group, council. You may need to check with the legal staff what is acceptable. The main purpose for the group is to be a functional group where trust grows over time to make data governance decisions together in the open space and not under the table and offline conversations. A well functioning group will take years to build. It will go through forming, storming, norming phases before it gets to a performing phase.

I found that educational sessions on small topics and examples of data use presented by actual data owners or users is the most effective way to start having more complex conversations. Storytelling is an absolute must. In government, this usually involves targeted assistance, training or services to a vulnerable group of people. Over time, hearing more examples and seeing results of data use brings more trust into this shared data governance exercise.

Technology tools will not change data governance in your organisation no matter how persuasive those sales people are. Last year, the U.S. Office of the Comptroller of the Currency (part of the U.S. Treasury) levied a $400 million civil money penalty against Citibank for not having a data governance programme . The financial industry is not short of resources to invest in various technology tools, but there are no shortcuts in having repeated conversations about data governance and establishing a formal programme. A strong data governance programme is the foundation for further work on building a culture of data-driven decision making. Communication is key. Regardless of how many phones, Zoom meetings, chats, emails and text messaging options we have, talking to people and listening to their concerns is the hardest part and requires no technology, just undivided presence in the conversation and patience. -Milda Aksamitauskas

__We want to hear what you think about this article. Pitch an article to us or send your comments to discussion@apolitical.co __

(Picture Credit: Pexels)