Four Steps Enterprises Can Take to Create a Data-Driven Work Culture

(Rawpixel.com/Shutterstock)

In the latest installment of the annual NewVantage Partners Big Data and AI Executive Survey of blue-chip organizations, it was discovered that:

• Only 31% of the organizations surveyed have succeeded in creating a data-driven enterprise
• 1% say business adoption of big data and AI initiatives remains a major challenge
• 95% blame company culture

Was I surprised to see these stats? No. Still, it’s incredibly disheartening all the same when all the ingredients you need to drive  big-data success are there for the taking. What’s really holding organizations back is the ability to put those all those ingredients together and implement them into the enterprise’s culture.

With a data-driven work culture at an organization’s core and permeated into all job functions, organizing, cleaning and defining data becomes part of the company’s philosophy and mission.

The NewVantage Partners survey found only 28.3% of organizations have been able to forge a data culture. Here are five steps businesses can take to become more data-minded:

1. Create a “Blessed” Source of High-Quality, Searchable Data

Data experts Carl Anderson and Michael Li once published a great piece on the overarching culture that’s required to create a data-minded organization. Anderson and Li talked about the need to create a single “blessed” source of clean, high-quality data that the entire company has access to.

After years of trying to create a single repository of data (enterprise data warehouses, data lakes, data hubs and so forth), many enterprises came to the conclusion that instead of shuffling data around, they are better off designating “blessed” copies of data and directing users to those copies through enterprise catalogs. In other words, they need to make data findable and provision it on demand, either in place or at the designated data source.

2. Make That Data Accessible

It’s not enough to simply identify high-quality data. People need to be able to find it, understand it and access it. To find it, a business glossary with business terms needs to be defined and data has to be tagged with those terms in a catalog, so analysts can use the terms they are familiar with to find and understand the data.

Once data is found, it needs to be provisioned or made available to the user either in place or as a copy. This can be a tall order in the era of increased regulation and the need for airtight governance. You can’t just sit and wait for IT to get around to responding to a request for access. That access has to be easily requested and provided according to pre-defined user roles.

Furthermore, while there are all kinds of great visualization and other tools for analytics that democratize data for everyone within the organization, these users still need to be reasonably data literate to come up with their insights and present their findings in an effective manner. Those findings then need to be factored into the decision-making process.

3. Make Data Easier to Share

Evidence-based decision making requires the sharing of data. Unfortunately, we usually run into the data access paradox. The team that needs data does not know what data is available, and the team that has the data does not know what data is needed.

To connect the dots, you need to help teams share metadata about their data. Since technical metadata can be cryptic or ambiguous, the best way to make sure that other teams can find and understand each other’s data is to use commonly understood business terms or tags to describe it.

This renders the data findable and understandable. It also eliminates another reason data often gets stuck in silos—namely, the team that needs the data does not understand it, so they rely on the team that “owns” it to allocate resources and schedule time to convert it into the shape the requesting team needs. By rendering the data comprehensible and documenting it using commonly understood business terms, the team that owns the data is no longer required to spend time and resources to help other team.

4. Encourage Data-Driven Decision Making

Even after making data findable and usable, organizations often fail in the final step of getting workers to use it. Some data-driven enterprises do not allow any product to be shipped without a measurement place. This includes both the plan that justifies the business case and the ongoing measurements that ensure that the product performs as expected and the use case is satisfied.

If that’s too ambitious, at least start the journey by asking for data-driven justification for business decisions. It must be communicated at all levels that gut-driven decision making is out and data-driven decision making is in. Data needs to permeate decision making from the top down, throughout every department and every business unit. Leadership needs to lead both by example and through a mandate that the burden of proof always comes down to the actual proof. (An anecdote is not research!) Policies and processes must be developed to keep teams honest.

It is also essential that people be thoroughly trained. This is critical, because even well-intentioned workers can be stubbornly resistant to behavioral change even when it’s simply a matter of getting them to use a new tool or adopt an official taxonomy.

Once you’re able to put all these elements in place, your organization can start benefiting from the more data-minded culture that’s been established. And then, the NewVantage Partners survey stat that shows most companies aren’t yet data-driven (and won’t be anytime soon) will be good news, because your organization will be.

About the author: Alex Gorelik is author of O’Reilly Media’s “The Enterprise Big Data Lake: Delivering the Promise of Big Data and Data Science,” and the founder and CTO of data cataloging company Waterline Data. Prior to Waterline Data, Gorelik served as senior vice president and general manager of Informatica’s Data Quality Business Unit, driving R&D, product marketing and product management for an $80 million business. He joined Informatica from IBM, where he was an IBM Distinguished Engineer for the Infosphere team. IBM acquired Gorelik’s second startup, Exeros (now Infosphere Discovery), where he was founder, CTO and vice president of engineering. Previously, he was cofounder, CTO and vice president of engineering at Acta Technology, a pioneering ETL and EII company, which was subsequently acquired by Business Objects.

Related Items:

How to Build a Data-Driven Culture

Carts & Horses: Why You Need to Focus on Data First

Reshaping the Business with Imagination: Four Steps to Digital Transformation Success