AWS Cloud Enterprise Strategy Blog
How to Create a Data-Driven Culture
“If we have data, let’s look at data. If all we have are opinions, let’s go with mine.”
― Jim Barksdale
Conversations about data-driven enterprises often focus on tools, big data, and the technological advancements that have made storing, processing, and analyzing data faster and cheaper. While these are all important, creating a data-driven culture across the enterprise is essential to move beyond just a few successful data initiatives and islands of excellence limited to certain business areas.
A data-driven culture embraces the use of data in decision making. It treats data as a strategic asset of the company by making data widely available and accessible. It focuses on capturing, cleaning, and curating meaningful data from across the business. It promotes frequent experimentation to learn and improve. It realizes that a strong foundation of data is critical for differentiating the business through artificial intelligence (AI) and machine learning (ML). It is a culture with a high level of data literacy and a belief that data helps everyone perform better.
How can leaders create a data-driven culture?
Walk the Talk
To create a data-driven culture, executive sponsorship is necessary but not sufficient. The C-suite must go beyond just supporting it. They need to remain engaged and involved, visibly marrying data with good business instincts to make decisions. This includes relentlessly culling the binders of unnecessary reports, spreadsheets, and dashboards generated just to satisfy them and create the illusion of a data-driven culture. Data should be for a purpose—not the purpose.
Executives should also actively seek data that may be contrary to their beliefs. Changing the course when new data is available sends a powerful signal to the organization about leadership’s commitment to creating a data-driven culture.
Data is empowering, but it can also evoke strong emotions. Data-driven cultures bring transparency and accountability to the whole enterprise. This can sometimes be uncomfortable. Typical enterprises have groups whose primary responsibility is to provide data and insights for others to consume. They fear losing relevance or control of the narrative. The C-suite must step in to overcome the resistance and organizational inertia.
Organize for Success
Each data initiative should be managed like a product rather than a byproduct of something else. This means setting up the right organizational structure to support it. My colleague, Joe Chung, talks about creating an Analytics Center of Excellence in his eBook Becoming a Data-Driven Organization. I like to think about it in terms of what we call a single-threaded leader at Amazon: a fully empowered leader whose full-time job is to make that one thing happen. (You can read more about the single threaded leader from Dave Limp and Jeff Wilke in the Forbes story here.) To create a data-driven culture, make sure there is an empowered leader who is waking up and worrying about just that one thing.
There is an increasing trend of breaking out data engineering and analytics from IT, which sometimes creates friction. Regardless of the reporting structure, IT should play an important role beyond just creating technical capabilities to support data initiatives. IT is in a unique position to have a complete view of the end-to-end business cycle, cross-departmental workflows, and transactional systems that hold a lot of useful insights. Before I became CTO, I led the global products, applications, and data team. Having an end-to-end ownership enabled my team to not only create a better data analytics platform that broke down silos, but also to address gaps in transactional systems to capture and utilize data better. Create a structure that is fully accountable but is also inclusive to avoid creating more fences to go around.
Treat Data as an Organizational Asset, Not a Departmental Property
In many enterprises, data silos are guarded zealously by organizational silos. Often, it is coming from a meaningful place, that of ensuring that the rest of the organization understands the context, variables, calculations, and patterns of certain data elements to use them the right way. Do not let this be yet another reason to deprive the enterprise of access to data, but rather convert these guardians into educators. Let these departmental data experts take ownership of teaching and helping others in the company use data the right way and increase the data literacy of the enterprise.
These silos also inhibit data correlation. Organizations often look at revenue, cost, inventory, and customer feedback in isolation because of how sales, finance, operation, and customer support each guard their own set of data. In today’s digital world, organizations must be more dynamic in finding and applying different levers using correlations to enhance customer experience and create better products using data. Create a light-weight data governance structure that starts with the goal of enabling more access to data, not restricting data.
Democratize the Data
A data-driven culture is not just about using data to make big decisions. A data-driven culture enables and empowers a lot of small, daily decisions by frontline employees using the data. In the digital economy, speed matters and using data to quickly test product ideas, design decisions, and hypotheses can increase business agility. High-frequency enterprises that deliver business value frequently move away from HiPPO (highest paid person’s opinion) based decision-making to a decentralized data-driven decision-making process.
Take the example of Epic Games and their massively successful game Fortnite, which attracts more than 125 million players globally. Epic uses AWS to gain an up-to-the-minute understanding of gamers’ satisfaction and interaction, and then they make this data available to their game designers. The game designers use this data to make decisions about fresh new experiences to add, like autogenerating maps that feel handcrafted for each game or deciding which weapons to introduce or discontinue. This creates a better experience for the gamers and a more engaged user community.
Speak the Same Language
Language has played an essential role since ancient times in establishing and maintaining cultures. A common language helps communicate the values, beliefs, and ideas that form the culture. Data-driven culture is no different. To create a data-driven culture, enterprises need to create a common vocabulary around data. This starts with the definition of key business metrics that the whole company understands, but then goes beyond to also identify variables that feed into those metrics. This is much more difficult than it sounds, because functions within an enterprise are often measured based on different success metrics.
To bring consistency, create a small set of key outcomes that the whole enterprise is accountable for and has visibility into. Then break them down into smaller metrics by functions that directly feed into the overall outcome. Do not simply measure what you can; measure what you should. The key is to keep this list of metrics short. Once you’ve identified them, agree on a common definition, and ensure that everyone understands it. It is important to create ongoing mechanisms to ensure consistency, accuracy, and education around them.
In summary, a data-driven culture thrives when the senior leadership is engaged, middle-management is empowered, frontline employees are energized, and silos are eliminated. Data-driven culture is a fun environment to work in, as it brings objectivity, transparency, and innovation. By driving a data-driven culture at scale, successful organizations turn data into a differentiator in the marketplace and a unifier within their company.
—Ishit
Twitter | LinkedIn | Blogs | Email
How to Build Data Capabilities, Ishit Vachhrajani
In Search of Silver Bullets: Moving Beyond Dreaming of Data, Phil Le-Brun
The CFO as Catalyst for the Data-Driven Enterprise, AWS Executive Insights
The Power of the Data-Driven Enterprise, AWS Executive Insights