How Data-Driven Organizations Overcome Culture and Technical Barriers

Experienced technology and data leaders shared their experiences establishing data-driven organizations during a recent Coffee with Digital Trailblazers episode. I host the event Fridays at 11 am ET, and while we’ve discussed many leadership topics tied to digital transformation over the weeks, this was our first episode where we focused on data practices.

Citizen data practices, proactive data governance, and delivering value from machine learning are key themes in my writing and speaking. I wrote about my experiences of being buried in bad data in Digital Trailblazer and shared many of my best practices in Driving Digital. My recent posts include how to expand citizen data science from successful analysts to data-driven organizations and on eradicating “IT Business Alignment” by empowering a data-driven partnership. Data-driven organizational practices are a major focus area for StarCIO’s Center of Excellence programs.

So, I was anxious to hear how they would respond to my questions:

1. What is your definition of a data-driven organization? What does it look like? What aspects are truly “critical”?
2. People-first approaches: What cultural barriers did you overcome and how?
3. Tech debt, data debt, security, governance: How can leaders unpack all the barriers and deliver value faster?

5 key insights for CIOs, chief data officers, and data leaders

Our discussions tend to bounce around people, process, and technology concepts during the coffee hour. It makes for a fast and interesting discussion, but capturing the key takeaways isn’t easy. This episode was filled with insights, and here are key recommendations from the group on what data-driven organizations do well:

Focus on value – Start with the end in mind and review the business value first and use these goals as a guide to trace back to relevant data sources. Once at the sources, you can then perform a more diligent review around data quality, processing issues, data governance gaps, and data management issues through a prioritized lens of what’s required to deliver accurate and timely business outcomes.

Get the language right – There’s enough data jargon and technologies to make a business leader’s head spin. Avoid asking them to assume the role of the data owner and define data quality KPIs; instead, get into the weeds by reviewing facts, data definitions, and actionable metrics. You can’t jump into insights without reviewing the data, and business leaders can best understand data quality issues when examples illustrate them.

Develop disciplined data scientists – Leaders advised many of the best practices: (1) Draft a vision statement before building anything, (2) Apply agile data practices, and (3) Document and share the data lineage, calculations, and machine learning models, used in any dashboards, analytics, and models.

Build trust by maturing data practices – They treat data as a corporate asset, implying that data must deliver value but requires ongoing investment. They focus on “democratizing data” by making the data easy to work with for everyone in the organization. “I shouldn’t need to consult with a subject matter expert to access or understand the data.” They also recognize that a handful of internal data products (dashboards, models, etc.) are far more useful than maintaining hundreds of customized reports.

Hire storytelling Digital Trailblazers – While many enterprises are in a race to hire top Ph.D. data scientists versed in the latest ML and AI capabilities, leaders of data-driven organizations are focused on the ability to influence and drive smarter and faster decision-making. Dashboards and models are tools, and the key skill is translating insights into stories that influence leaders to look beyond the status quo and develop transformational mindsets.

Key data practices for Digital Trailblazers

After the coffee hour, I followed up with Joanne Friedman, PhD., CEO, and principal of smart manufacturing at Connektedminds, to detail some of her insights. “In digital transformation or industry 4.0, the flexibility of data choreography means we can optimize the value of the data by placing it in different contexts,” she says.

Data choreography was a new term for me, so I asked Joanne to share its definition and relate it to data lineage and data provenance. Here’s her reply:

• Data provenance is data’s historical record keeper and is responsible for providing a list of origins, including inputs, entities, systems, and processes related to specific data.
• Data lineage provides an in-depth description of where data comes from, including its analytic life cycle.
• Data choreography is a design pattern used in software engineering where each component or service interacts with other components or services through an exchange of messages.

These definitions made me think of recipes in a cookbook: (i) Data provenance is the ingredients, where they’re sourced, and their overall quality, (ii) data lineage is the steps to follow in the recipe, and (iii) data choreography speaks to how easy and repeatable the recipe is in creating the dish.

The ability to focus on value, adopt data-driven practices, and use storytelling to influence decision-makers is the choreography required for data-driven organizations to deliver transformative results.

I hope you’ll join us for future episodes of Coffee with Digital Trailblazers!